No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Feature Selection And Classification Methodology For The Detection of Knee-Joint Disorders
Abstract
Vibroarthographic (VAG) signals emitted from the knee joint disorder provides an early diagnostic tool. The nonstationary and nonlinear nature of VAG signal makes an important aspect for feature extraction. In this work, we investigate VAG signals by proposing a wavelet based decomposition. The VAG signals are decomposed into sub-band signals of different frequencies. Nonlinear features such as recurrence quantification analysis (RQA), approximate entropy (ApEn) and sample entropy (SampEn) are extracted as features of VAG signal. A total of twenty four features form a vector to characterize a VAG signal. Two feature selection (FS) techniques, apriori algorithm and genetic algorithm (GA) selects six and four features as the most significant features. Least square support vector machines (LS-SVM) and random forest are proposed as classifiers to evaluate the performance of FS techniques. Results indicate that the classification accuracy was more prominent with features selected from FS algorithms. Results convey that LS-SVM using the apriori algorithm gives the highest accuracy of 94.31% with false discovery rate (FDR) of 0.0892. The proposed work also provided better classification accuracy than those reported in the previous studies which gave an accuracy of 88%. This work can enhance the performance of existing technology for accurately distinguishing normal and abnormal VAG signals. And the proposed methodology could provide an effective non-invasive diagnostic tool for knee joint disorders.
... Features related to the time domain were extracted by Nalband et al. [125] and include recurrence quantification analysis (RQA), approximate entropy (ApEn), and sample entropy (SampEn). Recurrence quantification analysis (RQA) is used to analyse complex, nonlinear dynamic systems. ...
... The lowest ACC of 74.19% was achieved for the following features: K, S, CWT 100-200 Hz, and CWT 200-500 Hz; whereas, the highest ACC of 86.77% was achieved for the feature set consisting of K, S, CWT 100-200 Hz, CWT 200-500 Hz, CWT 500-1000 Hz, and CWT 1000-2500 Hz. A higher classification result using the LS-SVM method was achieved by Nalband et al. [125]. An ACC of 91.01% was obtained using the genetic algorithm (four features), the use of all 24 features produced an ACC of 93.13%, while the use of the a priori algorithm resulted in an increase in ACC to 94.31%. ...
... Presentation of the accuracy level of the classification methods used in the selected works[76,77,85,87,88,92,100,104,108,119,121,125]. ...
The ageing population and the resulting number of physical and health problems are now a major social and economic challenge around the world. Osteoarthritis is a common disease among older people. It can affect any joint, but it most often affects the knee, hip, and hand joints. Osteoarthritis of the knee joint significantly affects everyday life, limiting daily activities. Patients affected by this disease face many ailments, such as pain, stiffness, and a reduced of range of joint motion. In order to implement quick and effective treatment and prevent the development of the disease, accurate and early diagnosis is important. This will contribute to prolonging the health of the joints. Available methods for diagnosing osteoarthritis include conventional radiography, MRI, and ultrasound, but these methods are not suitable for screening. Over the years, there have been proposals to use vibroarthrography as a new, cheap, and noninvasive screening method for cartilage damage. The paper reviews recent studies on vibroarthrography as a diagnostic method for knee osteoarthritis. The aim of the study is to organise the current knowledge regarding the diagnosis of osteoarthritis of the knee joint and vibroarthrography as a proposal for a new diagnostic method.
... Nalband et al. [15] investigated the Vibroarthographic signals by introducing a decomposition approach for decomposed into many sub-band signals with various frequencies. The existing feature selection algorithms, namely apriori and genetic algorithms, select 4 and 6 features as most contributed features. ...
... Here, the risk level is zero represents as (0,0,0,0), the risk level is very low represents as (0, 15,20), the risk level is low indicates the values as (15,30,45), the risk is medium level represents as (35,50,65), the risk level is high than it represents with the range of values as (55,70, 85) and the risk is extremely high which indicates with the values (80, 85, 100, 100). Moreover, the defuzzification is a final stage which provides the disease level as risk in the form of numeric value into linguistic value. ...
... Here, the risk level is zero represents as (0,0,0,0), the risk level is very low represents as (0, 15,20), the risk level is low indicates the values as (15,30,45), the risk is medium level represents as (35,50,65), the risk level is high than it represents with the range of values as (55,70, 85) and the risk is extremely high which indicates with the values (80, 85, 100, 100). Moreover, the defuzzification is a final stage which provides the disease level as risk in the form of numeric value into linguistic value. ...
This work proposes a new intelligent disease prediction system for predicting the disease and also knowing the current status of the dead diseases such as diabetic, heart and cancer diseases. More number of people are affecting and losing their life early due to these diseases so that these are also called as dead diseases. The proposed disease prediction and monitoring system consists of two phases such as feature selection and classification phases. In the feature selection phase, a newly proposed feature selection algorithm called conditional random field and mutual information-based feature selection algorithm is used for identifying the most contributed features that are used to enhance the prediction accuracy. In the classification phase, a newly proposed fuzzy-aware multilayer backpropagation neural network is applied for predicting and monitoring the diabetic disease and heart disease effectively. Here, newly generated fuzzy rules are also incorporated for making effective decision on patient records. The proposed prediction and monitoring system is used to predict and monitor the heart, diabetic and cancer diseases. The experiments have been conducted for evaluating the performance of the proposed disease prediction and monitoring system by using UCI Machine Learning Repository datasets and also proved that as better than the existing disease prediction systems in terms of precision, recall, F-measure and prediction accuracy.
... There is a number of specific measures, that can be extracted from the RP, which are collectively called the Recurrence Quantification Analysis (RQA). Some of those feature were previously used in a number of VAG-related studies [5,37,43,96,158]. ...
... Gong et al. [41] used SampEn with parameters m = 2 and r = 0.2. The same parameters were used in study by Nalband et al. [158] to obtain both ApEn and SampEn. Moreover, Kręcisz and Bączkowicz [5] used Mulsiscale Sample Entropy, i.e., the SampEn obtained for multiple time scale signals. ...
The impact of overload and a sedentary lifestyle on the health of knee joints can be linked to the joint structure’s reliance on the forces acting on its surface and on regular movement. Both excessive loading of the joint and absence of physical activity can lead to the cartilage degradation and reduction in the amount and quality of synovial fluid. Minor damages to the cartilage that may not cause pain, can contribute to more severe knee joint conditions. Vibroarthrography (VAG), an imaging method, assesses joint function quality by measuring vibrations generated during motion, producing a VAG signal. However, despite significant body of VAG-related studies, there is still a lack of consensus regarding some of the specific methods and features extracted from the VAG signal.
This dissertation presents an extensive examination of VAG signals derived from five classes of knee joint conditions. These include three Chondromalacia Patellae stages, one Osteoarthritis class, and a control group consisting of healthy knee joints. The thesis of this work proposes that the digital signal processing methods, in the context of VAG signal analysis, specifically in the time domain, frequency domain, and time-frequency domain, will enable extraction of features that lead to a classification accuracy higher than the current state-of-the-art (0.69). It was achieved through a thorough analysis of extracted signal features, optimization of parametric features, and selection of the most informative feature set. Ultimately, the most accurate classifier proved to be Linear Support Vector Machine trained on 110 features, achieving an accuracy of 0.80.
... Therefore, traditional signal processing techniques could not accurately investigate VAG signals. Reported literature has been comforted with methods based on time and frequency domain [6][7][8][9]. The frequency and time domain represent the time-frequency analysis of the input data/signal which describes the time varying frequency content. ...
Background: This paper presents AUTOENCODE-KNEE, a novel approach for automatic feature extraction from the time-frequency distribution of knee joint signals. Knee joint signals often contain valuable information crucial for diagnosing various musculoskeletal disorders. However, manually extracting relevant features from these signals can be time-consuming and subjective. Method: To address this challenge, we propose utilizing a convolutional neural network (CNN)-based autoencoder architecture for automatic feature extraction. The autoencoder is trained on a dataset comprising time-frequency representations of knee joint signals, learning to encode and decode the input signals while preserving important features. By leveraging the inherent ability of CNNs to capture spatial dependencies, the autoencoder effectively learns to extract discriminative features from the complex time-frequency domain. Result: Our experimental results demonstrate the efficacy of AUTOENCODE-KNEE in automatically extracting meaningful features from knee joint signals. We evaluate the extracted features on various classification tasks related to musculoskeletal disorder diagnosis, showcasing the utility of the proposed approach in aiding healthcare professionals in accurate and efficient diagnosis. Conclusion: In summary, AUTOENCODE-KNEE offers a promising solution for automatic feature extraction from knee joint signals, potentially revolutionizing how musculoskeletal disorders are diagnosed and treated.
... In this study, least square support vector machines signal processing was introduced. However, this study was also performed on small sample (n = 24) [53]. Other approaches towards signal processing have also been published. ...
... 5.5.1. Feature selection aims at determining strongly relevant, weakly relevant, irrelevant, and redundant features (Jovic et al., 2015;Nalband et al., 2016). It maximizes the relevance in data by rejecting the irrelevant and redundant leather image features. ...
... Another strategy to calculate the fluctuating features had been proposed (Yang et al. 2014) where two important features scaling and envelope are considered for analysis, and the classification of VAG signals is performed using LS-SVM. Different frequency ranges are calculated for VAG signals, and the frequency-domain features are composed (Nalband 2016) where the feature selection is done through GA and apriori algorithm, and the classification is performed through LS-SVM. Binary classification of VAG signals (i.e., normal and abnormal) had been performed (Saif Nalband et al. 2018) where time-frequency-based features are obtained using feature extraction methods, and the classification is performed using LS-SVM classifier. ...
Analysis of disorders in the human body using computational methods had given greater impact for diagnosis and cure. Vibroarthrographic signals are noninvasive modality that considers vibrations that are obtained from the human knee joints as a signal and analyzes the stability of the human knee joint. The signal analysis reaches the complete classification phase using the machine learning strategies in which feature engineering plays a vital role. The clinical cloud is used by the medical organizations to maintain the patient’s record, and the sample data are transferred to the specific diagnosis device where there is a need for cloud security measures. In this article, VAG signals are transferred from the model clinical cloud using security measures and analyzed by considering the number of feature selection and feature extraction strategies which are further justified by the classification results obtained through the machine learning algorithms. This security-enhanced research paves way for the clinicians and researchers to choose the appropriate security measures, feature analysis, and classifiers to obtain better results. A benchmark dataset that has 89 VAG signals are utilized for constructing the feature strategies, and the results are discussed.
... VAG is based on the analysis of vibroacoustic signals produced by the friction created inside the joint. Previous studies regarding knee joints showed that OA knees have a greater frequency, higher peaks, and longer duration concerning the vibroacoustic emissions compared to healthy knees [6,34,35]. An up-to-date VAG is not only used to determine the condition of the cartilage, but also for other intra-articular elements of the joints, such as the menisci in the knee joint and the temporomandibular joint's disc [5,36]. ...
Current research concerning the repeatability of the joint’s sounds examination in the temporomandibular joints (TMJ) is inconclusive; thus, the aim of this study was to investigate the repeatability of the specific features of the vibroarthrogram (VAG) in the TMJ using accelerometers. The joint sounds of both TMJs were measured with VAG accelerometers in two groups, study and control, each consisting of 47 participants (n = 94). Two VAG recording sessions consisted of 10 jaw open/close cycles guided by a metronome. The intraclass correlation coefficient (ICC) was calculated for seven VAG signal features. Additionally, a k-nearest-neighbors (KNN) classifier was defined and compared with a state-of-the-art method (joint vibration analysis (JVA) decision tree). ICC indicated excellent (for the integral below 300 Hz feature), good (total integral, integral above 300 Hz, and median frequency features), moderate (integral below to integral above 300 Hz ratio feature) and
poor (peak amplitude feature) reliability. The accuracy scores for the KNN classifier (up to 0.81) were higher than those for the JVA decision tree (up to 0.60). The results of this study could open up a new field of research focused on the features of the vibroarthrogram in the context of the TMJ, further improving the diagnosing process.
... Henceforth, there is a need to propose an expert framework that demonstrates the danger of disease and furthermore supports to find the arrangements utilizing information. The assessment paper we center around winning methodologies for diabetes recognition in order to know the ongoing advances in the field of diabetes under human services [8]. In the wake of examining every one of the strategies used to analyze diabetes disease, we make still some exertion in diagnosing the disease with high rightness. ...
At last decade, the development of diverse models and the excessive data creation leads to an enormous production of dataset and source. The healthcare field offers rich in information and it needs to be analyzed to identify the patterns present in the data. The commonly available massive amount of healthcare data characterizes a rich data field. The way of extracting the medical design is difficult because of the characteristics of healthcare data like massive, real, and complicated details. Various machine learning (ML) algorithms has developed to predict the existence of the diabetes disease. Due to the massive quantity of diabetes disease dataset, clustering techniques can be applied to group the data before classifying it. A new automated clustering based classification model is applied for the identification of diabetes. To cluster the healthcare data, sequential clustering (SC) model is applied. Then, logistic regression (LR) model is applied for the effective categorization of the clustered data. The experimentations have been directed by the benchmark dataset. The simulation outcomes demonstrate that the efficiency of the SC-LR method beats the prevailing methods to predict the diabetes diseases.
... The processing stage is probably the most complex of three stages since it could theoretically involve an infinite number of steps to drive its accuracy to 100%. Today's processing algorithms consist of many steps including feature selection [26], statistical training and retraining, network layering, and evaluation. ...
Computer-aided diagnosis (CAD) plays a key role in automating and enhancing the diagnosis of complex neurological disorders. Computers are not just used to automate the final diagnosis step but for the design of sensors, the preprocessing unit, and the processing unit as well. Today, it is an essential requirement that these CAD systems have low latency and low power consumption, which is not possible using a traditional software-based system that utilizes a complex processor to execute an operating system that supports the diagnostic application. The processor is clocked at a high frequency to execute all such supplementary processes, which increases the power consumption of a software-based system. Moreover, software-based systems inherently have high latency. Hardware accelerators overcome these limitations. Hardware acceleration is supported by the advent of the field-programmable gate array (FPGA), allowing rapid testing and deployment of hardware systems. Moreover, the novel neuromorphic platform has the potential to efficiently accelerate neural networks as well, which was not possible with the FPGA.In this chapter, we conduct a systematic review of hardware-accelerated neurodiagnostic systems. Various preprocessing, feature extraction, and diagnostic accelerators are studied. The articles are evaluated according to the implemented algorithm, computational complexity, latency, power/resource consumption, and accuracy. 99 articles are evaluated from a corpus of 1273 articles extracted from the most popular libraries. The review highlights the trade-offs associated with hardware-accelerated CAD systems. The reviewed literature is presented as a catalog enabling researchers and designers to make well-informed decisions when implementing custom CAD systems. This is the first such comprehensive review in our knowledge.KeywordsHardware accelerationField-programmable gate arrayNeurological diagnosisSystematic literature reviewEncephalograms
... It generated an impressive accuracy of 92.38%. Authors in [12] designed a predictive framework for the prediction of knee joint risks, and they used VAC signals for the purpose. Attribute selectors used were the apriori method and genetic search. ...
Advanced predictive analytics coupled with an effective attribute selection method plays a pivotal role in the precise assessment of chronic disorder risks in patients. Traditional attribute selection approaches suffer from premature convergence, high complexity, and computational cost. On the contrary, heuristic-based optimization to supervised methods minimizes the computational cost by eliminating outlier attributes. In this study, a novel buffer-enabled heuristic, a memory-based metaheuristic attribute selection (MMAS) model, is proposed, which performs a local neighborhood search for optimizing chronic disorders data. It is further filtered with unsupervised K-means clustering to remove outliers. The resultant data are input to the Naive Bayes classifier to determine chronic disease risks’ presence. Heart disease, breast cancer, diabetes, and hepatitis are the datasets used in the research. Upon implementation of the model, a mean accuracy of 94.5% using MMAS was recorded and it dropped to 93.5% if clustering was not used. The average precision, recall, and F-score metric computed were 96.05%, 94.07%, and 95.06%, respectively. The model also has a least latency of 0.8 sec. Thus, it is demonstrated that chronic disease diagnosis can be significantly improved by heuristic-based attribute selection coupled with clustering followed by classification. It can be used to develop a decision support system to assist medical experts in the effective analysis of chronic diseases in a cost-effective manner.
... And finally, the last step is to compute the Eigen vectors and Eigen values of the covariance matrix obtained above [8]. These eigen values are sorted as The other blocks correspond to misclassification [23]. ...
A non-invasive technique using knee joint vibroarthographic (VAG) signals can be used for the early diagnosis of knee joint disorders. Among the algorithms devised for the detection of knee joint disorders using VAG signals, algorithms based on entropy measures can provide better performance. In this work, the VAG signal is preprocessed using wavelet decomposition into sub band signals. Features of the decomposed sub bands such as approximate entropy, sample entropy & wavelet energy are extracted as a quantified measure of complexity of the signal. A feature selection based on Principal Component Analysis (PCA) is performed in order to select the significant features. The extracted features are then used for classification of VAG signal into normal and abnormal VAG using support vector machine. It is observed that the classifier provides a better accuracy with feature selection using principal component analysis. And the results show that the classifier was able to classify the signal with an accuracy of 82.6%, error rate of 0.174, sensitivity of 1.0 and specificity of 0.888.
... Such systems allow to detect and identify abnormalities, helping doctors make accurate diagnoses and appropriate treatment (Anthimopoulos et al. 2016;H.-D. Cheng et al. 2003;Gonzalez-Diaz 2018;Nalband et al. 2016;Suzuki 2013;Verma et al. 2016;Yanase et al. 2019;Zhou et al. 2015). Image classification has been proving capable of providing valuable cancer-fighting benefits, by classifying for example breast lesions (Zhou et al. 2015) or skin lesions (N. ...
Image classification, a key research in image processing and artificial intelligence, is of fundamental importance for an intelligent system to exploit and manage efficiently the visual information. The objective is to develop algorithms that automatically find the category, to which an image sample belongs, given training samples. In our studies, we focus on the research and applications of sparse representation based algorithms for image classification including but not limited to faces, objects and skin lesions. A key emphasis of this study is to formulate the sparse representation-based classification problems in specific domains, like wavelet and quaternion wavelet, in order to enhance classes separation performance. Further, our goal is to implement the novel method to computer-assisted melanoma diagnosing, performed on dermoscopic images. Melanoma is the most deadly type of skin cancer. Fortunately, skin lesions are curable if they are diagnosed and treated early enough. Due to this reason, the automated computer-assisted melanoma diagnosing has attracted great interest to researchers nowadays.
... A model called nine decision trees is developed, and it provides better results when compared to the decision tree and bragging algorithm. The main focus is the algorithm that are used for classification and the measuring the performance of DT algorithm and NB algorithm using the accuracy prediction technique and they proved that both algorithms provides better results after testing and in addition to that they stated that the decision tree is more cost effective than naïve Bayes when the dataset has less attributes and instance [17,22]. ...
Heart disease is a complex disease that affects a large number of people worldwide. The timely and accurate detection of heart disease is critical in healthcare, particularly in the field of cardiology. In this article, we proposed a system for diagnosing heart disease that is both efficient and accurate, and it is based on machine-learning techniques. The diagnosis of heart disease is found to be a serious concern, so the diagnosis has to be done remotely and regularly to take the prior action. In the present world, finding the prevalence of heart disease has become a key research area for the researchers and many models have crown proposed in the recent year. The optimization algorithm plays a vital role in heart disease diagnosis with high accuracy. Important goal of this work is to develop a hybrid GCSA which represents a genetic-based crow search algorithm for feature selection and classification using deep convolution neural networks. From the obtained results, the proposed model GCSA shows increase in the classification accuracy by obtaining more than 94% when compared to the other feature selection methods.
... Next, the selected optimal features are fed to radial basis function for classifying the medical data. S. Nalband, et al, [51] used a genetic algorithm and apriori algorithm to select the significant features from the extracted feature vectors. Then, random forest and Least Square SVM (LS-SVM) classification techniques are applied to precisely distinguish the abnormal and normal vibroarthographic signals. ...
The early diagnosis of chronic diseases plays a vital role in the field of healthcare communities and biomedical, where it is necessary for detecting the disease at an initial phase to reduce the death rate. This paper investigates the use of feature selection, dimensionality reduction and classification techniques to predict and diagnose the chronic disease. The appropriate selection of attributes plays a crucial role in improving the classification accuracy of the diagnosis systems. Additionally, dimensionality reduction techniques effectively improve the overall performance of the machine learning algorithms. On chronic disease databases, the classification techniques deliver efficient predictive results by developing intelligent, adaptive and automated system. Parallel and adaptive classification techniques are also analyzed in chronic disease diagnosis which is used to stimulate the classification procedure and to improve the computational cost and time. This survey article represents the overview of feature selection, dimensionality reduction and classification techniques and their inherent benefits and drawbacks.
... A model called nine decision trees is developed, it provides better results when compared to the decision tree and bragging algorithm. The main focus is the algorithm that are used for classification and the measuring the performance of DT algorithm and NB algorithm using the accuracy prediction technique and they proved that both algorithms provides better results after testing and in addition to that they stated that the decision tree is more cost effective than naïve Bayes when the dataset has less attributes and instance (Nalband et al. 2016;Saxena et al. 2015). ...
The diagnosis of heart disease is found to be a serious concern, so the diagnosis has to be done remotely and regularly to take the prior action. In the present world finding the prevalence of heart disease has become a key research area for the researchers and many models have been proposed in the recent year. The optimization algorithm plays a vital role in heart disease diagnosis with high accuracy. Important goal of this work is to develop a hybrid GA-ABC which represents a genetic based artificial bee colony algorithm for feature-selection and classification using classifier ensemble techniques. The ensemble classifier consists of four algorithms like support vector machine, random forest, Naïve Bayes, and decision tree. From the obtained results, the proposed model GA-ABC-EL shows increase in the classification accuracy by obtaining more than 90% when compared to the other feature selection methods.
... Wu et al. [73] used an SVM based on the entropy and envelope amplitude features and achieved an overall accuracy of 83.56%. Nalband et al. [86] utilised an a priori algorithm with DOI: http://dx.doi.org/10.5772/intechopen.92868 a least-square SVM classifier and claim accuracy of 94.31% with a false discovery rate of 0.0892. Kręcisz [87] achieved accuracies of >90% using a logistic regressionbased method. ...
The joints of the human body, especially the knees, are continually exposed to varying loads as a person goes about their day. These loads may contribute to damage to tissues including cartilage and the development of degenerative medical conditions such as osteoarthritis (OA). The most commonly used method currently for classifying the severity of knee OA is the Kellgren and Lawrence system, whereby a grade (a KL score) from 0 to 4 is determined based on the radiographic evidence. However, radiography cannot directly depict cartilage damage, and there is low inter-observer precision with this method. As such, there has been a significant activity to find non-invasive and radiation-free methods to quantify OA, in order to facilitate the diagnosis and the appropriate course of medical action and to validate the development of therapies in a research or clinical setting. A number of different teams have noted that variation in knee joint sounds during different loading conditions may be indicative of structural changes within the knee potentially linked to OA. Here we will review the use of acoustic methods, such as acoustic Emission (AE) and vibroarthrography (VAG), developed for the monitoring of knee OA, with a focus on the issues surrounding data collection and analysis.
... Also, Bastos and Caiado [9] have used RP to determine trend on stock markets, while Addo et al. [4] also used RP for financial applications. Furthermore, Rashvandi and Nasrabadi [46] have been able to distinguish between different breath sounds using RQA and Nalband et al. [41] determined features to classify knee disorders using cross recurrence plots (CRP). ...
Dealing with electroencephalogram signals (EEG) is often not easy. The lack of predicability and complexity of such non-stationary, noisy and high-dimensional signals is challenging. Cross recurrence plots (CRP) have been used extensively to deal with the detection of subtle changes in signals, even when the noise is embedded in the signal. In this contribution, a total of 121 children performed visual attention experiments and a proposed methodology using CRP and a Welch Power Spectral Distribution have been used to classify then between those who have ADHD and the control group. Additional tools were presented to determine to which extent this methodology is able to classify accurately and avoid misclassifications, thus demonstrating that this methodology is feasible to classify EEG signals from subjects with ADHD. The experimental results indicate that the proposed methodology shows higher accuracy in comparison with methods proposed by other authors, providing that the correct recurrence tools are selected. Also, this methodology does not require extensive training such as the methods proposed using classical machine learning algorithms. Furthermore, this proposed methodology shows that it is not required to manually discriminate events among the EEG electrodes since CRP can detect even the smallest changes in the signal even when it has embedded noise. Lastly, the results were compared with baseline machine learning methods to prove experimentally that this methodology is consistent and the results repeatable. Given the right CRP metrics, an accuracy of up to 97.25% was achieved, indicating that this methodology outperformed many of the state-of-the-art techniques.
... Often, some variables contain redundant or irrelevant data, leading to poor classification results [41,42]. To overcome this issue, feature selection methods have been used in computer-assisted diagnosis to select the most relevant features while the classifier's performance increases [43][44][45]. ...
Rheumatoid arthritis (RA) is an autoimmune disorder that typically affects people between 23 and 60 years old causing chronic synovial inflammation, symmetrical polyarthritis, destruction of large and small joints, and chronic disability. Clinical diagnosis of RA is stablished by current ACR-EULAR criteria, and it is crucial for starting conventional therapy in order to minimize damage progression. The 2010 ACR-EULAR criteria include the presence of swollen joints, elevated levels of rheumatoid factor or anti-citrullinated protein antibodies (ACPA), elevated acute phase reactant, and duration of symptoms. In this paper, a computer-aided system for helping in the RA diagnosis, based on quantitative and easy-to-acquire variables, is presented. The participants in this study were all female, grouped into two classes: class I, patients diagnosed with RA (n = 100), and class II corresponding to controls without RA (n = 100). The novel approach is constituted by the acquisition of thermal and RGB images, recording their hand grip strength or gripping force. The weight, height, and age were also obtained from all participants. The color layout descriptors (CLD) were obtained from each image for having a compact representation. After, a wrapper forward selection method in a range of classification algorithms included in WEKA was performed. In the feature selection process, variables such as hand images, grip force, and age were found relevant, whereas weight and height did not provide important information to the classification. Our system obtains an AUC ROC curve greater than 0.94 for both thermal and RGB images using the RandomForest classifier. Thirty-eight subjects were considered for an external test in order to evaluate and validate the model implementation. In this test, an accuracy of 94.7% was obtained using RGB images; the confusion matrix revealed our system provides a correct diagnosis for all participants and failed in only two of them (5.3%).
Graphical abstract
... This method is novel as it localizes upon a generic triboacsoutic signal which is emitted by the problematic joint during motion [26]. These signals are known to be quasi-stationary or non -stationary [5], [20]. The reason it was designed to localize upon generic triboacoustic signals was to make it robust, since the types of injuries, individuals, and joints are unique. ...
... accuracy was obtained with this hybrid model. A predictive model for knee joint disorder detection using VAC signals was proposed by [28]. The a priori algorithm and the genetic algorithm were used as feature evaluators while LS-SVM were the classifiers used for the study. ...
There is a consistent rise in chronic diseases worldwide. These diseases decrease immunity and the quality of daily life. The treatment of these disorders is a challenging task for medical professionals. Dimensionality reduction techniques make it possible to handle big data samples, providing decision support in relation to chronic diseases. These datasets contain a series of symptoms that are used in disease prediction. The presence of redundant and irrelevant symptoms in the datasets should be identified and removed using feature selection techniques to improve classification accuracy. Therefore, the main contribution of this paper is a comparative analysis of the impact of wrapper and filter selection methods on classification performance. The filter methods that have been considered include the Correlation Feature Selection (CFS) method, the Information Gain (IG) method and the Chi-Square (CS) method. The wrapper methods that have been considered include the Best First Search (BFS) method, the Linear Forward Selection (LFS) method and the Greedy Step Wise Search (GSS) method. A Decision Tree algorithm has been used as a classifier for this analysis and is implemented through the WEKA tool. An attribute significance analysis has been performed on the diabetes, breast cancer and heart disease datasets used in the study. It was observed that the CFS method outperformed other filter methods concerning the accuracy rate and execution time. The accuracy rate using the CFS method on the datasets for heart disease, diabetes, breast cancer was 93.8%, 89.5% and 96.8% respectively. Moreover, latency delays of 1.08 s, 1.02 s and 1.01 s were noted using the same method for the respective datasets. Among wrapper methods, BFS' performance was impressive in comparison to other methods. Maximum accuracy of 94.7%, 95.8% and 96.8% were achieved on the datasets for heart disease, diabetes and breast cancer respectively. Latency delays of 1.42 s, 1.44 s and 132 s were recorded using the same method for the respective datasets. On the basis of the obtained result, a new hybrid Attribute Evaluator method has been proposed which effectively integrates enhanced K-Means clustering with the CFS filter method and the BFS wrapper method. Furthermore, the hybrid method was evaluated with an improved decision tree classifier. The improved decision tree classifier combined clustering with classification. It was validated on 14 different chronic disease datasets and its performance was recorded. A very optimal and consistent classification performance was observed. The mean values for accuracy, specificity, sensitivity and f-score metrics were 96.7%, 96.5%, 95.6% and 96.2% respectively.
... It has been previously shown that VAG possess not only high accuracy and specificity when differentiating synovial joints' deteriorations with various biomechanical and morphological origins, but also is sensitive for identifying the changes in arthrokinematics related to the level of the joint load at the end of a performed task [16][17][18][19][20]. From an arthrology perspective, this finding seems to be particularly important, because the load on articular surfaces is one of the most essential factors affecting the level of kinetic friction and joint wear [7,21]. ...
Background:
The patellofemoral joint (PFJ) provides extremely low kinetic friction, which results in optimal arthrokinematic motion quality. Previous research showed that these friction-reducing properties may be diminished due to the increase in articular contact forces. However, this phenomenon has not been analyzed in vivo during functional daily-living activities. The aim of this study was the vibroarthrographic assessment of changes in PFJ arthrokinematics during squats with variated loads.
Methods:
114 knees from 57 asymptomatic subjects (23 females and 34 males) whose ages ranged from 19 to 26 years were enrolled in this study. Participants were asked to perform 3 trials: 4 repetitions of bodyweight squats (L0), 4 repetitions of 10 kg barbell back loaded squats (L10), 4 repetitions of 20 kg barbell back loaded squats (L20). During the unloaded and loaded (L10, L20) squats, vibroarthrographic signals were collected using an accelerometer placed on the patella and were described by the following parameters: variation of mean square (VMS), mean range (R4), and power spectral density for frequency of 50-250 Hz (P1) and 250-450 Hz (P2).
Results:
Obtained results showed that the lowest values were noted in the unloaded condition and that the increased applied loads had a significant concomitant increase in all the aforementioned parameters bilaterally (p < 0.05).
Conclusion:
This phenomenon indicates that the application of increasing knee loads during squats corresponds to higher intensity of vibroacoustic emission, which might be related to higher contact stress and kinetic friction as well as diminished arthrokinematic motion quality.
... It has been previously shown that VAG possess not only high accuracy and speci city when differentiating synovial joints' deteriorations with various biomechanical and morphological origins, but also is sensitive for identifying the changes in arthrokinematics related to the level of the joint load at the end of a performed task [16][17][18][19][20]. From an arthrology perspective, this nding seems to be particularly important, because the load on articular surfaces is one of the most essential factors affecting the level of kinetic friction and joint wear [7,21]. ...
Background: The patellofemoral joint (PFJ) provides extremely low kinetic friction, which results in optimal arthrokinematic motion quality. Previous research showed that these friction-reducing properties may be diminished due to the increase in articular contact forces. However, this phenomenon has not been analyzed in vivo during functional daily-living activities. The aim of this study was the vibroarthrographic assessment of changes in PFJ arthrokinematics during squats with variated loads.
Methods: 114 knees from 57 asymptomatic subjects (23 females and 34 males) whose ages ranged from 19 to 26 years were enrolled in this study. Participants were asked to perform
3 trials: 4 repetitions of bodyweight squats (L0), 4 repetitions of 10 kg barbell back loaded squats (L10), 4 repetitions of 20 kg barbell back loaded squats (L20). During the unloaded and loaded (L10, L20) squats, vibroarthrographic signals were collected using an accelerometer placed on the patella and were described by the following parameters: variation of mean square (VMS), mean range (R4), and power spectral density for frequency of 50-250 Hz (P1) and 250-450 Hz (P2).
Results: Obtained results showed that the lowest values were noted in the unloaded condition and that the increased applied loads had a significant concomitant increase in all the aforementioned parameters bilaterally (p<0.05).
Conclusion: This phenomenon indicates that the application of increasing knee loads during squats corresponds to higher intensity of vibroacoustic emission, which might be related to higher contact stress and kinetic friction as well as diminished arthrokinematic motion quality.
... It has been previously shown that VAG possess not only high accuracy and speci city when differentiating synovial joints' deteriorations with various biomechanical and morphological origins, but also is sensitive for identifying the changes in arthrokinematics related to the level of the joint load at the end of a performed task [16][17][18][19][20]. From an arthrology perspective, this nding seems to be particularly important, because the load on articular surfaces is one of the most essential factors affecting the level of kinetic friction and joint wear [7,21]. ...
Background: The patellofemoral joint (PFJ) provides extremely low kinetic friction, which results in optimal arthrokinematic motion quality. Previous research showed that these friction-reducing properties may be diminished due to the increase in articular contact forces. However, this phenomenon has not been analyzed in vivo during functional daily-living activities. The aim of this study was the vibroarthrographic assessment of changes in PFJ arthrokinematics during squats with variated loads.
Methods: 114 knees from 57 asymptomatic subjects (23 females and 34 males) whose ages ranged from 19 to 26 years were enrolled in this study. Participants were asked to perform 3 trials: 4 repetitions of bodyweight squats (L0), 4 repetitions of 10 kg barbell back loaded squats (L10), 4 repetitions of 20 kg barbell back loaded squats (L20). During the unloaded and loaded (L10, L20) squats, vibroarthrographic signals were collected using an accelerometer placed on the patella and were described by the following parameters: variation of mean square (VMS), mean range (R4), and power spectral density for frequency of 50-250 Hz (P1) and 250-450 Hz (P2).
Results: Obtained results showed that the lowest values were noted in the unloaded condition and that the increased applied loads had a significant concomitant increase in all the aforementioned parameters bilaterally (p<0.05).
Conclusion: This phenomenon indicates that the application of increasing knee loads during squats corresponds to higher intensity of vibroacoustic emission, which might be related to higher contact stress and kinetic friction as well as diminished arthrokinematic motion quality.
... There are other methods to accomplish feature selection which are detailed elsewhere. [77][78][79] Deep learning (DL) is a type of technique for ML in which higher levels of information are extracted using multiple levels, or layers, to transform data into more abstract, composite representations. 80 A major advantage for ML and DL approaches is that they can handle massive amounts of data and factors, making analytical progress a realistic goal for articular cartilage pathology risk assessment, staging and predicting progression of disease, and informing treatment decision making. ...
There is a critical unmet need in the clinical implementation of valid preventative and therapeutic strategies for patients with articular cartilage pathology based on the significant gap in understanding of the relationships between diagnostic data, disease progression, patient-related variables, and symptoms. In this article, the current state of classification and categorization for articular cartilage pathology is discussed with particular focus on machine learning methods and the authors propose a bedside–bench–bedside approach with highly quantitative techniques as a solution to these hurdles. Leveraging computational learning with available data toward articular cartilage pathology patient phenotyping holds promise for clinical research and will likely be an important tool to identify translational solutions into evidence-based clinical applications to benefit patients. Recommendations for successful implementation of these approaches include using standardized definitions of articular cartilage, to include characterization of depth, size, location, and number; using measurements that minimize subjectivity or validated patient-reported outcome measures; considering not just the articular cartilage pathology but the whole joint, and the patient perception and perspective. Application of this approach through a multistep process by a multidisciplinary team of clinicians and scientists holds promise for validating disease mechanism-based phenotypes toward clinically relevant understanding of articular cartilage pathology for evidence-based application to orthopaedic practice.
... It has been previously shown that VAG possess not only high accuracy and specificity when differentiating synovial joints' deteriorations with various biomechanical and morphological origins, but also is sensitive for identifying the changes in arthrokinematics related to the level of the joint load at the end of a performed task [16][17][18][19][20]. From an arthrology perspective, this finding seems to be particularly important, because the load on articular surfaces is one of the most essential factors affecting the level of kinetic friction and joint wear [7,21]. ...
Background: The patellofemoral joint (PFJ) provides extremely low kinetic friction, which results in optimal arthrokinematic motion quality. Previous research showed that these friction-reducing properties may be diminished due to the increase in articular contact forces. However, this phenomenon has been not analyzed in vivo during functional daily-living activities. The aim of this study was the vibroarthrographic assessment of changes in PFJ arthrokinematics during squats with variated loads.
Methods: Fifty-seven asymptomatic subjects were enrolled in this study. Participants were asked to perform 3 trials: 4 repetitions of bodyweight squats (L0), 4 repetitions of 10 kg barbell back loaded squats (L10), 4 repetitions of 20 kg barbell back loaded squats (L20). During the unloaded and loaded (L10, L20) squats, vibroarthrographic signals were collected using an accelerometer placed on the patella and were described by the following parameters: variation of mean square (VMS), mean range (R4), and power spectral density for frequency of 50-250 Hz (P1) and 250-450 Hz (P2).
Results: Obtained results showed that the lowest values were noted in the unloaded condition and that the increased applied loads had a significant concomitant increase in all the aforementioned parameters bilaterally (p<0.05).
Conclusion: This phenomenon indicates that the application of increasing knee loads during squats corresponds to higher intensity of vibroacoustic emission, which may result in higher contact stress and kinetic friction as well as diminished arthrokinematic motion quality.
Utilizing IoT devices for automated signal extraction and data processing is a cornerstone of Computer-Aided Diagnostics, addressing various clinical challenges. Among these, diagnosing osteoarthritis, a critical knee joint disorder early on is paramount to prevent severe joint damage. Vibroarthrography (VAG), a novel approach, leverages sound waves produced during knee joint movement to diagnose various stages of this disorder. This article presents a computational system based on VAG signals, seamlessly integrated with IoT devices for knee joint data extraction. Employing machine learning techniques facilitates the classification of osteoarthritis levels. By offering this system as consumer electronics, it reduces costs and radiation exposure compared to traditional clinical modalities. Our implementation gathered 187 clinical data points using the proposed computational system, integrating IoT devices to capture vibrations. Analyzing the recorded data involved computing various feature sets, enabling multiple classifications of osteoarthritis levels. Evaluation based on accuracy, precision, recall, and AUC demonstrated the efficacy of our proposed binary and multiclass classification models, indicating its potential as a mechanism for collecting and analyzing data for early-stage osteoarthritis detection.
Background and Purpose: We utilized the Perceived Racism Scale—Racism on the Job subscale—to assess how frequently Black nurses experienced racism on the job in the past year (ROTJ-Y) and throughout their lifetime (ROTJ-L). We aimed to assess the reliability and assess construct validity of each subscale in a sample of 53 nurses. Methods: Reliability was evaluated using Cronbach’s alphas, item correlations, and interitem correlations. Construct validity was examined using exploratory factor analysis. Results: Results demonstrated that the subscales are reliable and valid. Cronbach’s alphas for the ROTJ-Y and ROTJ-L were .93 and .91, respectively. Exploratory factor analysis revealed a unidimensional factor for both subscales. Conclusion: This study demonstrated that the Racism on the Job subscales are psychometrically sound measures of workplace racism among Black nurses.
Gene sequence classification can be seen as a challenging task due to the nonstationary, noisy and nonlinear characteristics of sequential data. The primary goal of this research is to develop a general solution approach for supervised DNA chromatogram (DNAC) classification in the absence of sufficient training data. Today, deep learning comes to the fore with its achievements, however this requires a lot of training data. Finding enough training data can be exceedingly challenging, particularly in the medical area and for rare disorders. In this paper, a novel supervised DNAC classification method is proposed, which combines three techniques to classify hepatitis virus DNA trace files as HBV and HCV. The features that are capable of reflecting the complex-structured sequential data are extracted based on both embedding and spectral entropies. After the supervised dimension reduction step, not only global behavior of the entropy features but also local behavior of the entropy features is taken into account for classification purpose. A memory-based learning, which cannot lose any information coming from training data as its nature, is being used as a classifier. Experimental results show that the proposed method achieves good results that although 19% training data is used, a performance of 92% is obtained.
Feature selection is a dimensionality reduction approach that aims to choose a small subset of the key characteristics from the original features by deleting redundant, superfluous, or noisy attributes. Improved learning performance, such as increased learning accuracy, lower processing costs, and better model interpretability, can all be attributed to feature selection. Without the proper utilization of FS approach, machine learning models can give inappropriate results. Researchers in computer vision, text mining, and other domains recently presented a variety of feature selection algorithms and shown their usefulness via theory and experiments. In this paper, work done by feature selection researchers to improve the performance of machine learning algorithms is presented which will be useful to budding researchers in this field.KeywordsMachine learningFeature selectionUnsupervisedFilterWrapperEmbedded
Chronic Kidney Disease (CKD) is a crucial life-threatening condition due to impaired kidney functionality and renal disease. In recent studies, Kidney disorder is considered one of the essential and deadliest issues that threaten patients’ survival with the lack of earlier prediction and classification. The earlier prediction process and the proper diagnosis help delay or stop the chronic disease progression into its final stage, where renal transplantation or dialysis is a known way of saving the patient’s life. Global studies reveal that nearly 10% of the population is affected by Chronic Kidney Disease (CKD), and millions die because of non-affordable treatment. Early detection of CKD from the biological parameters would save people from this crisis. Machine Learning algorithms are playing a predominant role in disease diagnosis and prognosis. This work generates compound features from CKD indicators by two novel algorithms: Correlation-based Weighted Compound Feature (CWCF) and Feature Significance based Weighted Compound Feature (FSWCF). Any learning algorithm is as good as its features. Hence, the features generated by these algorithms are validated on different machine learning algorithms as a test for generality. The simulation is done in MATLAB 2020a environment where various metrics like prediction accuracy gives superior results compared to multiple other approaches. The accuracy of CWCF over different methods like LR is 97.23%, Gaussian NB is 99%, SVM is 99.18%, and RF is 99.89%, which is substantially higher than the approaches without proper methods feature analysis. The results suggest that generated compound features improve the predictive power of the algorithms.
Objectives:
This study aimed to design a machine learning-based prediction framework to predict the presence or absence of systemic lupus erythematosus (SLE) in a cohort of Omani patients.
Methods:
Data of 219 patients from 2006 to 2019 were extracted from Sultan Qaboos University Hospital's electronic records. Among these, 138 patients had SLE, while the remaining 81 had other rheumatologic diseases. Clinical and demographic features were analysed to focus on the early stages of the disease. Recursive feature selection was implemented to choose the most informative features. The CatBoost classification algorithm was utilised to predict SLE, and the SHAP explainer algorithm was applied on top of the CatBoost model to provide individual prediction reasoning, which was then validated by rheumatologists.
Results:
CatBoost achieved an area under the receiver operating characteristic curve score of 0.95 and a sensitivity of 92%. The SHAP algorithm identified four clinical features (alopecia, renal disorders, acute cutaneous lupus and haemolytic anaemia) and the patient's age as having the greatest contribution to the prediction.
Conclusion:
An explainable framework to predict SLE in patients and provide reasoning for its prediction was designed and validated. This framework enables clinicians to implement early interventions that will lead to positive healthcare outcomes.
Knee Osteoarthritis (KOA) is a chronic joint disease characterized by degeneration of knee's articular cartilage. Comparing with the imaging diagnostic method, the vibroarthrographic (VAG) may be emerged as a new candidate for diagnosis of KOA in clinics. However, it is challenging for doctors to evaluate patients' condition by visually detecting VAGs due to the limited understanding about contained information. Besides, the basic physiological signals are also closely correlated with a greater risk of KOA. Based on this, we focus on studying computer-assisted diagnosis method of KOA (KOA-CAD) using multivariate information (i.e. VAGs and basic physiological signals) based on an improved deep learning model (DLM). Firstly, a new Laplace distribution-based strategy (LD-S) for classification in DLM is designed. Secondly, an aggregated multiscale dilated convolution network (AMD-CNN) is constructed to learn features from multivariate information of KOA patients. Then, a new KOA-CAD method is proposed by integrating the AMD-CNN with the LD-S to realize three CAD objectives, including the automatic KOA detection, the KOA early detection, and the KOA grading detection. Finally, the multivariate information collected clinically are applied to verify the proposed method, where the automatic detection accuracy, early detection accuracy and grading detection accuracy are 93.6%, 92.1%, and 84.2% respectively.
A person’s walking pattern can reveal important information about their health. Mounting multiple sensors onto loose clothing potentially offers a comfortable way of collecting data about walking and other human movement. This research investigates how well the data from three sensors mounted on the lateral side of clothing (on a pair of trousers near the waist, upper thigh and lower shank) correlate with the data from sensors mounted on the frontal side of the body. Data collected from three participants (two male, one female) for two days were analysed. Gait cycles were extracted based on features in the lower-shank accelerometry and analysed in terms of sensor-to-vertical angles (SVA). The correlations in SVA between the clothing- and body-mounted sensor pairs were analysed. Correlation coefficients above 0.76 were found for the waist sensor pairs, while the thigh and lower-shank sensor pairs had correlations above 0.90. The cyclical nature of gait cycles was evident in the clothing data, and it was possible to distinguish the stance and swing phases of walking based on features in the clothing data. Furthermore, simultaneously recording data from the waist, thigh, and shank was helpful in capturing the movement of the whole leg.
Background and Objective
Knee-joint vibroarthrographic (VAG) signal is an effective method for performing a non-invasive knee osteoarthritis (KOA) diagnosis, VAG signal analysis plays a crucial role in achieving the early pathological screening of the knee joint. In order to improve the accuracy of knee pathology screening and to investigate the method suitable for embedded in wearable diagnostic device for knee joint, this paper proposes a knee pathology screening method. Aiming to fill the gap of lacking suitable and unified evaluation indexes for single feature and fusion feature, this paper proposes feature separability evaluation criteria.
Methods
In this paper, we propose a knee joint pathology screening method based on feature fusion and dimension reduction combined with random forest classifier, as well as, the evaluation criteria of feature separability. As for pathological screening method, this paper proposes the idea of multi-dimensional feature fusion, using principal component analysis (PCA) to reduce the redundant part of fusion feature (F-F) to obtain deep fusion feature (D-F-F) with more separability. Meanwhile, this paper proposes the maximal information coefficient (MIC) and correlation matrix collinearity (CMC) feature evaluation criteria, these not only can be used as new feature quantitative metrics, but also illustrate that the divisibility of the deep fusion feature is more potent than that before feature dimension reduction.
Results
The experimental results show that the method in this paper has good performance in pathology classification on random forest classifier with 96% accuracy, especially the accuracy of SVM and K-NN are also improved after feature dimension reduction.
Conclusion
The results indicate that this classification research has high screening efficiency for KOA diagnosis and could provide a feasible method for computer-assisted non-invasive diagnosis of KOA. And we provide a novel way for separability evaluation of VAG signal features.
Oral glucose ingestion induces systemic changes of many blood metabolites related not only to glucose, but also other metabolites such as amino acids and lipids through many blood hormones. However, the detailed temporal changes in the concentrations of comprehensive metabolites and hormones over a long time by oral glucose ingestion are uncharacterized. We measured 83 metabolites and 7 hormones in 20 healthy human subjects in response to glucose ingestion. We characterized temporal patterns of blood molecules by four features: (i) the decomposability into “amplitude” and “rate” components, (ii) the similarity of temporal patterns among individuals, (iii) the relation of molecules over time among individuals, and (iv) the similarity of temporal patterns among molecules. Glucose and glucose metabolism-related hormones indicated a rapid increase, and citrulline and lipids, which indicated a rapid decrease, returned to fasting levels faster than amino acids. Compared to glucose metabolism-related molecules and lipids, amino acids showed similar temporal patterns among individuals. The four features of temporal patterns of blood molecules by oral glucose ingestion characterize the differences among individuals and among molecules.
Low energy availability (LEA) may persist in rugby players. However, timely assessment of energy balance is important but is difficult. Therefore, a practical index that reflects energy availability (EA) is essential. A total of 19 male college rugby players participated in a 2-week pre-season summer camp. Their blood sample was collected after overnight fast prior to (Pre), in the middle (Middle), and after (Post) the camp. Their physical activity in the first half of the camp was calculated using the additive factor method in the forwards (FW; numbers 1-8) and backs (BK; numbers 9-15). The participants were categorized as tight five (T5; numbers 1-5), back row (BR; numbers 6-8), and BK for analysis. All the participants lost weight during the camp (range: from -5.9% to -0.1%). Energy balance in the first half of the camp was negative. Transferrin saturation (TSAT) and serum iron levels significantly decreased to half, or even less, compared with the Pre levels at week 1 and remained low. The changes in TSAT and serum iron levels exhibited a significant positive correlation with the changes in body weight (R = 0.720; R = 0.627) and with energy intake (R = 0.410; R = 461) in T5. LEA occurs in rugby summer camp but is difficult to assess using weight change. Alternately, TSAT and serum iron levels after overnight fast may be better predictors of LEA.
The largest as well as complicated joint in the human body is knee joint. A minor pathology in human knee can lead to major discomfort to a person. The pathology may be from a simple stress in knee joints to the higher effective levels of pathology that includes osteoarthritis (OA), chondromalacia patella (CMP) and so on. Hence prediction as well as classification of knee joint pathology plays a vital role for computational medical diagnosis. Pathology identification can be categorized in to two levels invasive and non – invasive methodologies. Invasive are purely medical as well as painful process which are less preferred for predictions. Noninvasive are clinical and computational ones that includes X-Rays, Magnetic Resonance Imaging (MRI), Computed Tomography (CT), Vibroarthography (VAG) etc. This paper compares various classification methodologies utilized by Vibroarthographic (VAG) signal based knee joint disorder level identification and performs a binary classification of data samples as a survey.
It has been determined that variations in vibroarthrographic (VAG) signal characteristics have a direct association with various diseases of the knee joints including degenerative and traumatic etiology. Several studies have proposed temporal and time-frequency parameters for the analysis and classification of VAG signals; however, very few statistical modelling methods have been explored to analyze distinctions in the spectral domain characteristics of VAG signals. The aim of our work is to classify VAG signals into normal and abnormal thus portraying knee health condition by proposing a novel set of features. Spectral domain transformations of VAG signals were derived using the short-time Fourier transform (STFT) approach to extract statistical characteristic features from VAG signals. Several time-domain statistical features; the mean and variance values of spectral domain features namely – spectral mean, peak, skewness, kurtosis, flux, and slope were utilized to express VAG signal fluctuations. Further, mutual information test, ANOVA F-test, and Chi-square test were used to select significant features and assess their effectiveness. Random forest, Gaussian Naive Bayes, and support vector machine (SVM) algorithms were used to perform signal pattern classification. The results showed that subjects with knee joint abnormalities possess higher values of spectral flux, spectral slope, and shape factor than healthy subjects. An overall classification accuracy of 89.23%, sensitivity of 87.09% and specificity of 91.18% were obtained using the random forest classifier trained with features selected by the mutual information test. The proposed system could effectively screen almost 94% of subjects accurately based on area under the ROC curve.
Knee Osteoarthritis (KOA) is a common and chronic degenerative joint disease. Comparing with the traditional examinations (e.g., X-ray, MRI), the vibroarthrographic (VAG) examination, which is a low-costly, atraumatic, and at-home way, may open up new alternatives to KOA detection in clinic. However, it is a hard problem for doctors to evaluate the condition of KOA patients by visually detecting VAG signal due to the very limited understanding of pathological information included in VAG signals. Originated from this, we focus on exploring a reliable KOA auxiliary diagnostic method using VAG signals. In this paper, a new feature extraction method is first proposed, where the kernel-radius-based feature (KR-F) and statistic-based feature (S-F) are extracted respectively in the transient phase space of VAG signal. Furthermore, two features are integrated in the feature-fusion level (KR-S-FF), and then fed into the back propagation neural network (BPNN) to complete the KOA detection automatically. Finally, the proposed automatic KOA detection method is verified in a clinical VAG dataset, which is collected from one hospital in Xi’an, China. Simulation results convey that the proposed method gives the high detection accuracy of 98.2 % with sensitivity of 97.9 % and specificity of 98.5 %, showing that it may provide an effective non-invasive tool for KOA disorders.
Recently unique spans of genetic data are produced by researchers, there is a trend in genetic exploration using machine learning integrated analysis and virtual combination of adaptive data into the solution of classification problems. Detection of ailments and infections at early stage is of key concern and a huge challenge for researchers in the field of machine learning classification and bioinformatics. Considerate genes contributing to diseases are of huge dispute to a lot of researchers. This study reviews various works on Dimensionality reduction techniques for reducing sets of features that groups data effectively with less computational processing time and classification methods that contributes to the advances of RNA-Sequencing approach.
Dealing with electroencephalogram signals (EEG) are often not easy. The lack of predicability and complexity of such non-stationary, noisy and high dimensional signals is challenging. Cross Recurrence Plots (CRP) have been used extensively to deal with detecting subtle changes in signals even when the noise is embedded in the signal. In this contribution, a total of 121 children performed visual attention experiments and a proposed methodology using CRP and a Welch Power Spectral Distribution have been used to classify then between those who have ADHD and the control group. Additional tools were presented to determine to which extent the proposed methodology is able to classify accurately and avoid misclassifications, thus demonstrating that this methodology is feasible to classify EEG signals from subjects with ADHD. Lastly, the results were compared with a baseline machine learning method to prove experimentally that this methodology is consistent and the results repeatable.
This chapter has been inspired by the need to find an efficient method to study the changes in the cardiac autonomic regulation before and after smoking in young smokers. Appropriate processing and analysis of the RR intervals, extracted from the electrocardiogram (ECG) signals, is an important non-invasive tool to determine the autonomic nervous system (ANS)-based alteration of the cardiac physiology. The authors propose to analyze the RR interval signals by recurrence analysis and extracting the features thereof. In total, 17 young volunteers (smokers) participated in the study. The ECG signals were acquired before and after smoking. The statistical features obtained from the recurrence analysis were used as the input to artificial neural network (ANN)-based classification so as to classify the basal state from the post-stimulus state. A classification efficiency of 100% could be achieved using the recurrence features of the RR interval signals. Thus, the results evince that the proposed method is an efficient method to study the changes in the RR interval signal due to smoking.
We review accuracy estimation methods and compare the two most common methods: crossvalidation and bootstrap. Recent experimental results on arti cial data and theoretical results in restricted settings have shown that for selecting a good classi er from a set of classiers (model selection), ten-fold cross-validation may be better than the more expensive leaveone-out cross-validation. We report on a largescale experiment|over half a million runs of C4.5 and a Naive-Bayes algorithm|to estimate the e ects of di erent parameters on these algorithms on real-world datasets. For crossvalidation, we vary the number of folds and whether the folds are strati ed or not � for bootstrap, we vary the number of bootstrap samples. Our results indicate that for real-word datasets similar to ours, the best method to use for model selection is ten-fold strati ed cross validation, even if computation power allows using more folds. 1
We review several aspects of the analysis of time sequences, and concentrate on recent methods using concepts from the theory of nonlinear dynamical systems. In particular, we discuss problems in estimating attractor dimensions, entropies, and Lyapunov exponents, in reducing noise and in forecasting. For completeness and since we want to stress connections to more traditional (mostly spectrum-based) methods, we also give a short review of spectral methods.
A new graphical tool for measuring the time constancy of dynamical systems is presented and illustrated with typical examples.
Analysis of knee joint vibration (VAG) signals can provide quantitative indices for detection of knee joint pathology at an early stage. In addition to the statistical features developed in the related previous studies, we extracted two separable features, that is, the number of atoms derived from the wavelet matching pursuit decomposition and the number of significant signal turns
detected with the fixed threshold in the time domain. To perform a better classification over the data set of 89 VAG signals, we applied a novel classifier fusion system based on the dynamic weighted fusion (DWF) method to ameliorate the classification performance. For comparison, a single leastsquares support vector machine (LS-SVM) and the Bagging ensemble were used for the classification task as well. The results in terms of overall accuracy in percentage and area under the receiver operating characteristic curve obtained with the DWF-based classifier fusion method reached 88.76% and 0.9515, respectively, which demonstrated the effectiveness and superiority of the DWF method with two distinct features for the VAG signal analysis.
Epilepsy is one of the most common neurological disorders - approximately one in every 100 people worldwide are suffering from it. The electroencephalogram (EEG) is the most common source of information used to monitor, diagnose and manage neurological disorders related to epilepsy. Large amounts of data are produced by EEG monitoring devices, and analysis by visual inspection of long recordings of EEG in order to find traces of epilepsy is not routinely possible. Therefore, automated detection of epilepsy has been a goal of many researchers for a long time. This paper presents a novel method for automatic epileptic seizure detection. An optimized sample entropy (O-SampEn) algorithm is proposed and combined with extreme learning machine (ELM) to identify the EEG signals regarding the existence of seizure or not. To the knowledge of the authors, there exists no similar work in the literature. A public dataset was utilized for evaluating the proposed method. Results show that the proposed epilepsy detection approach achieves not only high detection accuracy but also a very fast computation speed, which demonstrates its huge potential for the real-time detection of epileptic seizures.
In this letter we discuss a least squares version for support vector machine (SVM) classifiers. Due to equality type constraints in the formulation, the solution follows from solving a set of linear equations, instead of quadratic programming for classical SVM''s. The approach is illustrated on a two-spiral benchmark classification problem.
This paper developed a diagnosis model based on Support Vector Machines (SVM) with a novel hybrid feature selection method to diagnose erythemato-squamous diseases. Our hybrid feature selection method, named IFSFFS (Improved F-score and Sequential Forward Floating Search), combines the advantages of filters and wrappers to select the optimal feature subset from the original feature set. In our IFSFFS, we firstly generalized the original F-score to the improved F-score measuring the discrimination of more than two sets of real numbers. Then we proposed to combine Sequential Forward Floating Search (SFFS) and our improved F-score to accomplish the optimal feature subset selection. Where, our improved F-score is an evaluation criterion for filters, while SFFS and SVM compose an evaluation system of wrappers. The best parameters of kernel function of SVM are found out by grid search technique with tenfold cross validation. Experiments have been conducted on five random training-test partitions of the erythemato-squamous diseases dataset from UCI machine learning database. The experimental results show that our SVM-based model with IFSFFS achieved the optimal classification accuracy with no more than 14 features as well.
This paper proposes three novel training methods, two of them based on the backpropagation approach and a third one based on information theory for multilayer perceptron (MLP) binary classifiers. Both backpropagation methods are based on the maximal-margin (MM) principle. The first one, based on the gradient descent with adaptive learning rate algorithm (GDX) and named maximum-margin GDX (MMGDX), directly increases the margin of the MLP output-layer hyperplane. The proposed method jointly optimizes both MLP layers in a single process, backpropagating the gradient of an MM-based objective function, through the output and hidden layers, in order to create a hidden-layer space that enables a higher margin for the output-layer hyperplane, avoiding the testing of many arbitrary kernels, as occurs in case of support vector machine (SVM) training. The proposed MM-based objective function aims to stretch out the margin to its limit. An objective function based on Lp-norm is also proposed in order to take into account the idea of support vectors, however, overcoming the complexity involved in solving a constrained optimization problem, usually in SVM training. In fact, all the training methods proposed in this paper have time and space complexities O(N) while usual SVM training methods have time complexity O(N (3)) and space complexity O(N (2)) , where N is the training-data-set size. The second approach, named minimization of interclass interference (MICI), has an objective function inspired on the Fisher discriminant analysis. Such algorithm aims to create an MLP hidden output where the patterns have a desirable statistical distribution. In both training methods, the maximum area under ROC curve (AUC) is applied as stop criterion. The third approach offers a robust training framework able to take the best of each proposed training method. The main idea is to compose a neural model by using neurons extracted from three other neural networks, each one previously trained by MICI, MMGDX, and Levenberg-Marquard (LM), respectively. The resulting neural network was named assembled neural network (ASNN). Benchmark data sets of real-world problems have been used in experiments that enable a comparison with other state-of-the-art classifiers. The results provide evidence of the effectiveness of our methods regarding accuracy, AUC, and balanced error rate.
Approximate entropy (ApEn), a recently developed mathematical formula quantifying regularity, was applied to fetal heart rate (FHR) data. Three groups were analyzed: 1) 19 women had normal labors (uncomplicated course of labor, vaginal delivery, no unusual FHR tracings, and 1- and 5-minute Apgar scores of at least 7 and 9, respectively; 2) 15 women had presumed fetal distress (severe cord or late decelerations, bradycardia, or tachycardia; delivery by cesarean with both arterial and venous cord pH above 7.20); and 3) 20 women had acidotic fetuses (both venous and arterial cord pH less than 7.20). Hourly mean (+/- SD) ApEn values for the three groups were: acidotic fetuses, 0.924 +/- 0.235, 102 hours; normal fetuses, 1.051 +/- 0.145, 97 hours; and nonacidotic "distressed" fetuses, 1.043 +/- 0.147, 74 hours. The ApEn values for nonacidotic, presumed distressed fetuses were not significantly different from those of normal fetuses (P greater than .75). Acidotic fetuses had many more instances of ApEn hourly values less than 0.8 (28%, 29 of 102) than did the normal and the nonacidotic, presumed distressed fetuses combined (5%, nine of 171). The probability that ApEn was less than 0.8 was larger for acidotic fetuses than for the other groups (P less than .00003), supporting the hypothesis that extremely regular FHR patterns imply a greater likelihood of acidosis. Significant hourly deviations in ApEn generally corresponded to drug administration or to physiologic changes such as cord compression and its relief. Thus ApEn, a major departure from variability statistics, appears to be able to detect subtle and possibly important differences in heart rate that are not visually apparent.
The detection and recording of vibration emission from human joints, a technique which we have termed "vibration arthrography", is a sensitive, non-invasive method for the objective study of the locomotor system. Using vibration sensors attached to bony prominences around the knee, we studied the joints of both normal and symptomatic subjects. Normal subjects produced three signal types--physiological patellofemoral crepitus, patellar clicks, and the lateral band signal. In symptomatic subjects we identified and categorised many signal types and related them to pathology. Lesions of the menisci produced distinctive signals, and it was possible not only to lateralise the tear, but in many cases to determine the type of meniscal injury present. Vibration arthrography promises to be a useful tool in the non-invasive diagnosis of knee disorders.
Physiological systems are best characterized as complex dynamical processes that are continuously subjected to and updated by nonlinear feedforward and feedback inputs. System outputs usually exhibit wide varieties of behaviors due to dynamical interactions between system components, external noise perturbations, and physiological state changes. Complicated interactions occur at a variety of hierarchial levels and involve a number of interacting variables, many of which are unavailable for experimental measurement. In this paper we illustrate how recurrence plots can take single physiological measurements, project them into multidimensional space by embedding procedures, and identify time correlations (recurrences) that are not apparent in the one-dimensional time series. We extend the original description of recurrence plots by computing an array of specific recurrence variables that quantify the deterministic structure and complexity of the plot. We then demonstrate how physiological states can be assessed by making repeated recurrence plot calculations within a window sliding down any physiological dynamic. Unlike other predominant time series techniques, recurrence plot analyses are not limited by data stationarity and size constraints. Pertinent physiological examples from respiratory and skeletal motor systems illustrate the utility of recurrence plots in the diagnosis of nonlinear systems. The methodology is fully applicable to any rhythmical system, whether it be mechanical, electrical, neural, hormonal, chemical, or even spacial.
Entropy, as it relates to dynamical systems, is the rate of information production. Methods for estimation of the entropy of a system represented by a time series are not, however, well suited to analysis of the short and noisy data sets encountered in cardiovascular and other biological studies. Pincus introduced approximate entropy (ApEn), a set of measures of system complexity closely related to entropy, which is easily applied to clinical cardiovascular and other time series. ApEn statistics, however, lead to inconsistent results. We have developed a new and related complexity measure, sample entropy (SampEn), and have compared ApEn and SampEn by using them to analyze sets of random numbers with known probabilistic character. We have also evaluated cross-ApEn and cross-SampEn, which use cardiovascular data sets to measure the similarity of two distinct time series. SampEn agreed with theory much more closely than ApEn over a broad range of conditions. The improved accuracy of SampEn statistics should make them useful in the study of experimental clinical cardiovascular and other biological time series.
Techniques to determine changing system complexity from data are evaluated. Convergence of a frequently used correlation dimension algorithm to a finite value does not necessarily imply an underlying deterministic model or chaos. Analysis of a recently developed family of formulas and statistics, approximate entropy (ApEn), suggests that ApEn can classify complex systems, given at least 1000 data values in diverse settings that include both deterministic chaotic and stochastic processes. The capability to discern changing complexity from such a relatively small amount of data holds promise for applications of ApEn in a variety of contexts.
Abnormal heart rate characteristics of reduced variability and transient decelerations are present early in the course of neonatal sepsis. To investigate the dynamics, we calculated sample entropy, a similar but less biased measure than the popular approximate entropy. Both calculate the probability that epochs of window length m that are similar within a tolerance r remain similar at the next point. We studied 89 consecutive admissions to a tertiary care neonatal intensive care unit, among whom there were 21 episodes of sepsis, and we performed numerical simulations. We addressed the fundamental issues of optimal selection of m and r and the impact of missing data. The major findings are that entropy falls before clinical signs of neonatal sepsis and that missing points are well tolerated. The major mechanism, surprisingly, is unrelated to the regularity of the data: entropy estimates inevitably fall in any record with spikes. We propose more informed selection of parameters and reexamination of studies where approximate entropy was interpreted solely as a regularity measure.
This book presents the cutting-edge technologies of knee joint vibroarthrographic signal analysis for the screening and detection of knee joint injuries. It describes a number of effective computer-aided methods for analysis of the nonlinear and nonstationary biomedical signals generated by complex physiological mechanics. This book also introduces several popular machine learning and pattern recognition algorithms for biomedical signal classifications. The book is well-suited for all researchers looking to better understand knee joint biomechanics and the advanced technology for vibration arthrometry. Dr. Yunfeng Wu is an Associate Professor at the School of Information Science and Technology, Xiamen University, Xiamen, Fujian, China.
The pathological condition in a degenerative knee joint may be assessed by analyzing the knee joint vibroarthrographic signals. With the severity level of the knee joint disorders evaluated by the computational methods, unnecessary imaging examination or open surgery can be prevented. In the present study, we used the k-nearest neighbor (k-NN) algorithm, a type of lazy learning approach, to classify the knee joint vibroarthrographic signals collected from healthy subjects and symptomatic patients with knee joint disorders. With the representative features of form factor and variance of the mean-square values, the k-NN algorithm is able to correctly discriminate 80% signals with the sensitivity of 0.71 and the specificity of 0.85, which is superior to the total accurate rate of 77% (sensitivity: 0.64, specificity: 0.85) provided by the Fisher's linear discriminant analysis.
In this chapter we consider bounds on the rate of uniform convergence. We consider upper bounds (there exist lower bounds as well (Vapnik and Chervonenkis, 1974); however, they are not as important for controlling the learning processes as the upper bounds).
Bagging predictors is a method for generating multiple versions of a predictor and using these to get an aggregated predictor. The aggregation averages over the versions when predicting a numerical outcome and does a plurality vote when predicting a class. The multiple versions are formed by making bootstrap replicates of the learning set and using these as new learning sets. Tests on real and simulated data sets using classification and regression trees and subset selection in linear regression show that bagging can give substantial gains in accuracy. The vital element is the instability of the prediction method. If perturbing the learning set can cause significant changes in the predictor constructed, then bagging can improve accuracy.
This article applies advanced signal processing and computational methods to study the subtle fluctuations in knee joint vibroarthrographic (VAG) signals. Two new features are extracted to characterize the fluctuations of VAG signals. The fractal scaling index parameter is computed using the detrended fluctuation analysis algorithm to describe the fluctuations associated with intrinsic correlations in the VAG signal. The averaged envelope amplitude feature measures the difference between the upper and lower envelopes averaged over an entire VAG signal. Statistical analysis with the Kolmogorov-Smirnov test indicates that both of the fractal scaling index (p=0.0001) and averaged envelope amplitude (p=0.0001) features are significantly different between the normal and pathological signal groups. The bivariate Gaussian kernels are utilized for modeling the densities of normal and pathological signals in the two-dimensional feature space. Based on the feature densities estimated, the Bayesian decision rule makes better signal classifications than the least-squares support vector machine, with the overall classification accuracy of 88% and the area of 0.957 under the receiver operating characteristic (ROC) curve. Such VAG signal classification results are better than those reported in the state-of-the-art literature. The fluctuation features of VAG signals developed in the present study can provide useful information on the pathological conditions of degenerative knee joints. Classification results demonstrate the effectiveness of the kernel feature density modeling method for computer-aided VAG signal analysis.
Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest. The generalization error for forests converges a.s. to a limit as the number of trees in the forest becomes large. The generalization error of a forest of tree classifiers depends on the strength of the individual trees in the forest and the correlation between them. Using a random selection of features to split each node yields error rates that compare favorably to Adaboost (Y. Freund & R. Schapire, Machine Learning: Proceedings of the Thirteenth International conference, ***, 148–156), but are more robust with respect to noise. Internal estimates monitor error, strength, and correlation and these are used to show the response to increasing the number of features used in the splitting. Internal estimates are also used to measure variable importance. These ideas are also applicable to regression.
Support vector machine (SVM) is a supervised machine learning approach that was recognized as a statistical learning apotheosis for the small-sample database. SVM has shown its excellent learning and generalization ability and has been extensively employed in many areas. This paper presents a performance analysis of six types of SVMs for the diagnosis of the classical Wisconsin breast cancer problem from a statistical point of view. The classification performance of standard SVM (St-SVM) is analyzed and compared with those of the other modified classifiers such as proximal support vector machine (PSVM) classifiers, Lagrangian support vector machines (LSVM), finite Newton method for Lagrangian support vector machine (NSVM), Linear programming support vector machines (LPSVM), and smooth support vector machine (SSVM). The experimental results reveal that these SVM classifiers achieve very fast, simple, and efficient breast cancer diagnosis. The training results indicated that LSVM has the lowest accuracy of 95.6107 %, while St-SVM performed better than other methods for all performance indices (accuracy = 97.71 %) and is closely followed by LPSVM (accuracy = 97.3282). However, in the validation phase, the overall accuracies of LPSVM achieved 97.1429 %, which was superior to LSVM (95.4286 %), SSVM (96.5714 %), PSVM (96 %), NSVM (96.5714 %), and St-SVM (94.86 %). Value of ROC and MCC for LPSVM achieved 0.9938 and 0.9369, respectively, which outperformed other classifiers. The results strongly suggest that LPSVM can aid in the diagnosis of breast cancer.
Diagnostic measures related to the deterioration of the articular cartilage surfaces of knee joints due to arthritis and other abnormalities may be derived from vibroarthrographic (VAG) signals. In the present work, we explore fractal analysis to parameterize the temporal and spectral variability of normal and abnormal VAG signals. The power spectrum analysis method was used with the 1/f model to derive estimates of the fractal dimension. Classification accuracy of up to Az = 0.74 was obtained, in terms of the area under the receiver operating characteristics curve, with a database of 89 VAG signals. The result compares well with the performance of other features derived in previous related works and could help in the detection and monitoring of knee-joint pathology.
The significance of detection and classification of power quality (PQ) events that disturbs the voltage and/or current waveforms in the electrical power distribution networks is well known. Consequently, in spite of a large number of research reports in this area, the problem of PQ event classification remains to be an important engineering problem. Several feature construction, pattern recognition, analysis, and classification methods were proposed for this purpose. In spite of the extensive number of such alternatives, a research on the comparison of “how useful these features with respect to each other using specific classifiers” was omitted. In this work, a thorough analysis is carried out regarding the classification strengths of an ensemble of celebrated features. The feature items were selected from well-known tools such as spectral information, wavelet extrema across several decomposition levels, and local statistical variations of the waveform. The tests are repeated for classification of several types of real-life data acquired during line-to-ground arcing faults and voltage sags due to the induction motor starting under different load conditions. In order to avoid specificity in classifier strength determination, eight different approaches are applied, including the computationally costly “exhaustive search” together with the leave-one-out technique. To further avoid specificity of the feature for a given classifier, two classifiers (Bayes and SVM) are tested. As a result of these analyses, the more useful set among a wider set of features for each classifier is obtained. It is observed that classification accuracy improves by eliminating relatively useless feature items for both classifiers. Furthermore, the feature selection results somewhat change according to the classifier used. This observation shows that when a new analysis tool or a feature is developed and claimed to perform “better” than another, one should always indicate the matching classifier for the feature because that feature may prove comparably inefficient with other classifiers.
This chapter reproduces the English translation by B. Seckler of the paper by Vapnik and Chervonenkis in which they gave proofs for the innovative results they had obtained in a draft form in July 1966 and announced in 1968 in their note in Soviet Mathematics Doklady. The paper was first published in Russian as ???????????? ??. ??. and ?????????????????????? ??. ??. ?? ?????????????????????? ???????????????????? ???????????? ?????????????????? ?????????????? ?? ???? ????????????????????????. ???????????? ???????????????????????? ?? ???? ???????????????????? 16(2), 264???279 (1971).
Approximate entropy (ApEn) and sample entropy (SampEn) are mathematical algorithms created to measure the repeatability or predictability within a time series. Both algorithms are extremely sensitive to their input parameters: m (length of the data segment being compared), r (similarity criterion), and N (length of data). There is no established consensus on parameter selection in short data sets, especially for biological data. Therefore, the purpose of this research was to examine the robustness of these two entropy algorithms by exploring the effect of changing parameter values on short data sets. Data with known theoretical entropy qualities as well as experimental data from both healthy young and older adults was utilized. Our results demonstrate that both ApEn and SampEn are extremely sensitive to parameter choices, especially for very short data sets, N ≤ 200. We suggest using N larger than 200, an m of 2 and examine several r values before selecting your parameters. Extreme caution should be used when choosing parameters for experimental studies with both algorithms. Based on our current findings, it appears that SampEn is more reliable for short data sets. SampEn was less sensitive to changes in data length and demonstrated fewer problems with relative consistency.
In epilepsy diagnosis or epileptic seizure detection, much effort has been focused on finding effective combination of feature extraction and classification methods. In this paper, we develop a wavelet-based sparse functional linear model for representation of EEG signals. The aim of this modeling approach is to capture discriminative random components of EEG signals using wavelet variances. To achieve this goal, a forward search algorithm is proposed for determination of an appropriate wavelet decomposition level. Two EEG databases from University of Bonn and University of Freiburg are used for illustration of applicability of the proposed method to both epilepsy diagnosis and epileptic seizure detection problems. For this data considered, we show that wavelet-based sparse functional linear model with a simple classifier such as 1-NN classification method leads to higher classification results than those obtained using other complicated methods such as support vector machine. This approach produces a 100 % classification accuracy for various classification tasks using the EEG database from University of Bonn, and outperforms many other state-of-the-art techniques. The proposed classification scheme leads to 99 % overall classification accuracy for the EEG data from University of Freiburg.
We introduce a novel wrapper Algorithm for Feature Selection, using Support Vector Machines with kernel functions. Our method is based on a sequential backward selection, using the number of errors in a validation subset as the measure to decide which feature to remove in each iteration. We compare our approach with other algorithms like a filter method or Recursive Feature Elimination SVM to demonstrate its effectiveness and efficiency.
In this paper, we developed a diagnosis model based on support vector machines (SVM) with a novel hybrid feature selection method to diagnose erythemato-squamous diseases. Our proposed hybrid feature selection method, named improved F-score and Sequential Forward Search (IFSFS), combines the advantages of filter and wrapper methods to select the optimal feature subset from the original feature set. In our IFSFS, we improved the original F-score from measuring the discrimination of two sets of real numbers to measuring the discrimination between more than two sets of real numbers. The improved F-score and Sequential Forward Search (SFS) are combined to find the optimal feature subset in the process of feature selection, where, the improved F-score is an evaluation criterion of filter method, and SFS is an evaluation system of wrapper method. The best parameters of kernel function of SVM are found out by grid search technique. Experiments have been conducted on different training-test partitions of the erythemato-squamous diseases dataset taken from UCI (University of California Irvine) machine learning database. Our experimental results show that the proposed SVM-based model with IFSFS achieves 98.61% classification accuracy and contains 21 features. With these results, we conclude our method is very promising compared to the previously reported results.
David Goldberg's Genetic Algorithms in Search, Optimization and Machine Learning is by far the bestselling introduction to genetic algorithms. Goldberg is one of the preeminent researchers in the field--he has published over 100 research articles on genetic algorithms and is a student of John Holland, the father of genetic algorithms--and his deep understanding of the material shines through. The book contains a complete listing of a simple genetic algorithm in Pascal, which C programmers can easily understand. The book covers all of the important topics in the field, including crossover, mutation, classifier systems, and fitness scaling, giving a novice with a computer science background enough information to implement a genetic algorithm and describe genetic algorithms to a friend.
Vibroarthrographic (VAG) signals, generated by human knee movement, are non-stationary and multi-component in nature and their time-frequency distribution (TFD) provides a powerful means to analyze such signals. The objective of this paper is to improve the classification accuracy of the features, obtained from the TFD of normal and abnormal VAG signals, using segmentation by the dynamic time warping (DTW) and denoising algorithm by the singular value decomposition (SVD). VAG and knee angle signals, recorded simultaneously during one flexion and one extension of the knee, were segmented and normalized at 0.5 Hz by the DTW method. Also, the noise within the TFD of the segmented VAG signals was reduced by the SVD algorithm, and a back-propagation neural network (BPNN) was used to classify the normal and abnormal VAG signals. The characteristic parameters of VAG signals consist of the energy, energy spread, frequency and frequency spread parameter extracted by the TFD. A total of 1408 segments (normal 1031, abnormal 377) were used for training and evaluating the BPNN. As a result, the average classification accuracy was 91.4 (standard deviation +/-1.7) %. The proposed method showed good potential for the non-invasive diagnosis and monitoring of joint disorders such as osteoarthritis.
The clinical performance of a laboratory test can be described in terms of diagnostic accuracy, or the ability to correctly classify subjects into clinically relevant subgroups. Diagnostic accuracy refers to the quality of the information provided by the classification device and should be distinguished from the usefulness, or actual practical value, of the information. Receiver-operating characteristic (ROC) plots provide a pure index of accuracy by demonstrating the limits of a test's ability to discriminate between alternative states of health over the complete spectrum of operating conditions. Furthermore, ROC plots occupy a central or unifying position in the process of assessing and using diagnostic tools. Once the plot is generated, a user can readily go on to many other activities such as performing quantitative ROC analysis and comparisons of tests, using likelihood ratio to revise the probability of disease in individual subjects, selecting decision thresholds, using logistic-regression analysis, using discriminant-function analysis, or incorporating the tool into a clinical strategy by using decision analysis.
We review the principles and practical application of receiver-operating characteristic (ROC) analysis for diagnostic tests. ROC analysis can be used for diagnostic tests with outcomes measured on ordinal, interval or ratio scales. The dependence of the diagnostic sensitivity and specificity on the selected cut-off value must be considered for a full test evaluation and for test comparison. All possible combinations of sensitivity and specificity that can be achieved by changing the test's cut-off value can be summarised using a single parameter; the area under the ROC curve. The ROC technique can also be used to optimise cut-off values with regard to a given prevalence in the target population and cost ratio of false-positive and false-negative results. However, plots of optimisation parameters against the selected cut-off value provide a more-direct method for cut-off selection. Candidates for such optimisation parameters are linear combinations of sensitivity and specificity (with weights selected to reflect the decision-making situation), odds ratio, chance-corrected measures of association (e. g. kappa) and likelihood ratios. We discuss some recent developments in ROC analysis, including meta-analysis of diagnostic tests, correlated ROC curves (paired-sample design) and chance- and prevalence-corrected ROC curves.
This article presents an automated, patient-specific method for the detection of epileptic seizure onset from noninvasive electroencephalography. We adopt a patient-specific approach to exploit the consistency of an individual patient's seizure and nonseizure electroencephalograms. Our method uses a wavelet decomposition to construct a feature vector that captures the morphology and spatial distribution of an electroencephalographic epoch, and then determines whether that vector is representative of a patient's seizure or nonseizure electroencephalogram using the support vector machine classification algorithm. Our completely automated method was tested on noninvasive electroencephalograms from 36 pediatric subjects suffering from a variety of seizure types. It detected 131 of 139 seizure events within 8.0+/-3.2 seconds of electrographic onset, and declared 15 false detections in 60 hours of clinical electroencephalography. Our patient-specific method can be used to initiate delay-sensitive clinical procedures following seizure onset, for example, the injection of a functional imaging radiotracer.
Sample entropy, a nonlinear signal processing approach, was used as a measure of signal complexity to evaluate the cyclic behavior of heart rate variability (HRV) in obstructive sleep apnea syndrome (OSAS). In a group of 10 normal and 25 OSA subjects, the sample entropy measure showed that normal subjects have significantly more complex HRV pattern than the OSA subjects (p < 0.005). When compared with spectral analysis in a minute-by-minute classification, sample entropy had an accuracy of 70.3% (69.5% sensitivity, 70.8% specificity) while the spectral analysis had an accuracy of 70.4% (71.3% sensitivity, 69.9% specificity). The combination of the two methods improved the accuracy to 72.9% (72.2% sensitivity, 73.3% specificity). The sample entropy approach does not show major improvement over the existing methods. In fact, its accuracy in detecting sleep apnea is relatively low in the well classified data of the physionet. Its main achievement however, is the simplicity of computation. Sample entropy and other nonlinear methods might be useful tools to detect apnea episodes during sleep.
Externally detected vibroarthrographic (VAG) signals bear diagnostic information related to the roughness, softening, breakdown, or the state of lubrication of the articular cartilage surfaces of the knee joint. Analysis of VAG signals could provide quantitative indices for noninvasive diagnosis of articular cartilage breakdown and staging of osteoarthritis. We propose the use of statistical parameters of VAG signals, including the form factor involving the variance of the signal and its derivatives, skewness, kurtosis, and entropy, to classify VAG signals as normal or abnormal. With a database of 89 VAG signals, screening efficiency of up to 0.82 was achieved, in terms of the area under the receiver operating characteristics curve, using a neural network classifier based on radial basis functions.
Decision trees are attractive classifiers due to their high
execution speed. But trees derived with traditional methods often cannot
be grown to arbitrary complexity for possible loss of generalization
accuracy on unseen data. The limitation on complexity usually means
suboptimal accuracy on training data. Following the principles of
stochastic modeling, we propose a method to construct tree-based
classifiers whose capacity can be arbitrarily expanded for increases in
accuracy for both training and unseen data. The essence of the method is
to build multiple trees in randomly selected subspaces of the feature
space. Trees in, different subspaces generalize their classification in
complementary ways, and their combined classification can be
monotonically improved. The validity of the method is demonstrated
through experiments on the recognition of handwritten digits
We consider the problem of discovering association rules between items in a large database of sales transactions. We presenttwo new algorithms for solving this problem that are fundamentally different from the known algorithms. Experiments with synthetic as well as real-life data show that these algorithms outperform the known algorithms by factors ranging from three for small problems to more than an order of magnitude for large problems. We also showhow the best features of the two proposed algorithms can be combined into a hybrid algorithm, called AprioriHybrid. Scale-up experiments show that AprioriHybrid scales linearly with the number of transactions. AprioriHybrid also has excellent scale-up properties with respect to the transaction size and the number of items in the database. 1 Introduction Database mining is motivated by the decision support problem faced by most large retail organizations [S + 93]. Progress in bar-code technology has made it possible for retail ...
In this paper, we examine a method for feature subset selection based on Information Theory. Initially, a framework for defining the theoretically optimal, but computationally intractable, method for feature subset selection is presented. We show that our goal should be to eliminate a feature if it gives us little or no additional information beyond that subsumed by the remaining features. In particular, this will be the case for both irrelevant and redundant features. We then give an efficient algorithm for feature selection which computes an approximation to the optimal feature selection criterion. The conditions under which the approximate algorithm is successful are examined. Empirical results are given on a number of data sets, showing that the algorithm effectively handles datasets with large numbers of features. 1 Introduction In the classic supervised learning task, we are given a training set of labeled fixed-length feature vectors, or instances, from which...
Z. Recurrence plots RPs often have fascinating structures, especially when the embedding dimension is 1. We identify four Z. basic patterns of a RP, namely, patterns along the main 458 diagonal, patterns along the 1358 diagonal, block-like structures, and square-like textures. We also study how the structures of and quantification statistics for RPs vary with the embedding parameters. By considering the distribution of the main diagonal line segments for chaotic systems, we relate some of the known statistics for the quantification of a RP to the Lyapunov exponent. This consideration enables us to introduce new ways of quantifying the diagonal line segments. Furthermore, we categorize recurrence points into two classes. A number of new quantities are identified which may be useful for the detection of nonstationarity in a time series, especially for the detection of a bifurcation sequence. A noisy transient Lorenz system is studied, to demonstrate how to identify a true bifurcation sequence, to interpret false bifurcation points, and to choose the embedding dimension. q 2000 Published by Elsevier Science B.V. All rights reserved.
- V Vigorita
- B Ghelman
- D Mintz
V. Vigorita, B. Ghelman, D. Mintz, Orthopaedic Pathology, M -Medicine Series, Lippincott Williams & Wilkins, 2008.
Fractal analysis of knee-joint vibroarthrographic signals via power spectral analysis
- R M Rangayyan
- F Oloumi
- Y Wu
- S Cai
R.M. Rangayyan, F. Oloumi, Y. Wu, S. Cai, Fractal analysis of
knee-joint vibroarthrographic signals via power spectral
analysis, Biomed. Signal Process. Control 8 (2013) 23-29,
http://dx.doi.org/10.1016/j.bspc.2012.05.004.