[Show abstract][Hide abstract] ABSTRACT: Sleep spindle is the one of important components determining N-REM (Non-Rapid Eye Movement) stage 2 in the sleep stages. The symptoms of N-REM stage 2 are sleep spindle and K-complex and here sleep spindles are automatically recognized by using time and frequency domain features belonging to EEG (Electroencephalograph) signals obtained from three patient subjects. In this study, the proposed method consists of two steps. In the first step, six time domain features have been extracted from raw EEG signals. As for the extraction of frequency domain features from raw EEG signals, Welch spectral analysis has been used and applied to raw EEG signals. By this way, 65 frequency domain features have been extracted and reduced from 65 to 4 features by using statistical measures including minimum, maximum, standard deviation, and mean values. Three feature sets including only time domain, only frequency domain, and both time and frequency domain features have been used and the numbers of these feature sets are 6, 4, and 10, respectively. In the second step, artificial neural network (ANN) with LM (Levenberg–Marquardt) has been used to classify the sleep spindles evaluated beforehand by sleep expert physicians. The obtained classification accuracies for three features sets in the classification of sleep spindles are 100%, 56.86%, and 93.84% by using LM-ANN (for ten node in hidden layer). The obtained results have presented that the proposed recognition system could be confidently used in the automatic classification of sleep spindles.
Expert Systems with Applications 03/2011; 38(3-38):2455-2461. DOI:10.1016/j.eswa.2010.08.034 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Sleep scoring is one of the most important methods for diagnosis in psychiatry and neurology. Sleep staging is a time consuming and difficult task conducted by sleep specialists. The purposes of this work are to automatic score the sleep stages and to help to sleep physicians on sleep stage scoring. In this work, a novel data preprocessing method called k-means clustering based feature weighting (KMCFW) has been proposed and combined with k-NN (k-nearest neighbor) and decision tree classifiers to classify the EEG (electroencephalogram) sleep into six sleep stages including awake, N-REM (non-rapid eye movement) stage 1, N-REM stage 2, N-REM stage 3, REM, and non-sleep (movement time). First of all, frequency domain features belonging to sleep EEG signal have been extracted using Welch spectral analysis method and composed 129 features from EEG signal relating each sleep stages. In order to decrease the features, the statistical features comprising minimum value, maximum value, standard deviation, and mean value have been used and then reduced from 129 to 4 features. In the second phase, the sleep stages dataset with four features has been weighted by means of k-means clustering based feature weighting. Finally, the weighted sleep stages have been automatically classified into six sleep stages using k-NN and C4.5 decision tree classifier. In the classification of sleep stages, the k values of 10, 20, 30, 40, 50, and 60 in k-NN classifier have been used and compared with each other. In the experimental results, while sleep stages has been classified with 55.88% success rate using k-NN classifier (for k value of 40), the weighted sleep stages with KMCFW has been recognized with 82.15% success rate k-NN classifier (for k value of 40). And also, we have investigated the relevance between sleep stages and frequency domain features belonging to EEG signal. These results have demonstrated that proposed weighting method have a considerable impact on automatic determining of sleep stages. This system could be used as an online system in the automatic scoring of sleep stages and helps to sleep physicians in the sleep scoring process.
Expert Systems with Applications 12/2010; 37(12-37):7922-7928. DOI:10.1016/j.eswa.2010.04.043 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Rheumatoid arthritis (RA) is a multi-systemic autoimmune disease that leads to substantial morbidity and mortality. In this paper, as spectral analysis methods of Multiple Signal Classification (MUSIC) method is used in order to extract the significant features from the right and left hand Ulnar artery Doppler signals for the diagnosis of RA disease. The MUSIC method has been used as subspace method. To extract features from Doppler signals obtained from the right and left hand Ulnar arterial the MUSIC method model degrees of 5, 10, 15, 20, and 25 were used. Then, an adaptive network based fuzzy inference system (ANFIS) was applied to features extracted from the right and left hand Ulnar artery Doppler signals for classifying RA disease. The methods are not new, but the study has a novelty in that the application area of these methods is new. In the hybrid model, the combination of MUSIC and ANFIS yielded classification accuracies of 95% (for a model degree of 20) using the right hand Ulnar artery and classification accuracies of 91.25% (for a model degree of 10) using left hand Ulnar artery Doppler signals in the diagnosis of RA disease. The proposed approach has potential to help with the early diagnosis of RA disease for the specialists who study this subject.
[Show abstract][Hide abstract] ABSTRACT: The cardiac, end-systolic and end-diastolic diameters values are very important m-mode cardiac parameters for infant, children, and adolescents, due to growing up body. These parameters, belonging to heart, must be known in order to make a decision about the subject. The expert decision occurs after comparing measured value to hard-copied charts. Hard-copied charts were prepared previously as a result of long statistical studies and these charts depend on a certain region.Our proposed method presents a valid virtual chart for the experts. The proposed method comprises of two stages: (i) data normalization based on euclidean distance (ii) normalized cardiac parameters predicting using adaptive neural fuzzy system. In order to present performance of the proposed method, mean absolute error, absolute deviation and two-fold cross-validation were used. In addition to performance criteria, different common normalization methods, z-score, decimal scaling and minimum–maximum normalization methods were used to compare.In this study, the aim is to create a valid virtual chart which helps the expert during making the decision about predicting end-systolic and end-diastolic cardiac m-mode values. The results were compared with real cardiac parameters by expert with 10years of medical experience.
Expert Systems with Applications 08/2010; 37(8):5720-5727. DOI:10.1016/j.eswa.2010.02.038 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: In this paper, a new feature selection named as multi-class f-score feature selection is proposed for sleep apnea classification having different disorder degrees (mild OSAS, moderate OSAS, serious OSAS, and non-OSAS). f-Score is used to measure the discriminating power of features in the classification of two-class pattern recognition problems. In order to apply the f-score feature selection to multi-class datasets, we have used the f-score feature selection as pairwise (in the form of two classes) in the diagnosis of obstructive sleep apnea syndrome (OSAS) with four classes. After feature selection process, MLPANN (Multi-layer perceptron artificial neural network) classifier is used to diagnose the OSAS having different disorder degrees. While MLPANN obtained 63.41% classification accuracy on the diagnosis of OSAS, the combination of MLPANN and multi-class f-score feature selection achieved 84.14% classification accuracy using 50–50% training–testing split of OSAS dataset with four classes. These results demonstrate that the proposed multi-class f-score feature selection method is effective and robust in determining the disorder degrees of OSAS.
Expert Systems with Applications 03/2010; 37(2):998-1004. DOI:10.1016/j.eswa.2009.05.075 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: In this study, Doppler signals received from the radial artery of the right hands of the 40 healthy subjects and 40 patients with rheumatoid arthritis were recorded. Some features of these signals were obtained using Subspace-based MUSIC method which is one of the spectral analysis methods, and the diseased cases were distinguished with Artificial Neural Networks classification method. In MUSIC method, 5,10,15,20 and 25 were used as model degrees. Test procedure was carried out after training with Artificial Neural Networks. Classification accuracy after the test results was 88.75% for 5 model degrees, 93.75% for 10 and 25 model degrees, 100% for 15 model degrees and 92.5% for 20 model degrees and all models had an average degree of 93.75% classification accuracy was obtained. The proposed approach has potential to help with the early diagnosis of RA disease for the specialists who study this subject.
[Show abstract][Hide abstract] ABSTRACT: Sleep staging has an effective role in diagnosing sleep disorders. Sleep staging is generally done by a sleep expert through examining Electroencephalogram (EEG), Electrooculogram (EOG), Electromyogram (EMG) signals of the patients and determining the stages of sleep in different time sections. This type of sleep staging is preferred among the sleep experts but because it is rather tiring and time consuming task, attention to the automatic sleep staging systems has been begun to increase. In this study, we obtained EEG, EMG and EOG signals of five healthy people in Meram Faculty of medicine to use in sleep staging and extracted 74 features from them. We analyzed the effects of these features on sleep staging. We utilized from the sequential feature selection algorithm and Artificial Neural Networks in this application. We determined which features are more effective in classification of sleep stages and by this way we tried to guide researchers who will use EEG, EMG and EOG features in sleep staging. The highest classification accuracy was obtained as 69.30% with use of four features.
[Show abstract][Hide abstract] ABSTRACT: All over the world, many portable devices need battery to run. Every expert has to use efficient hardware and software documentation to make battery last longer and make a correlation between microcontrollers’ duties and the remaining energy of batteries. In order to make battery last longer, battery information must be evaluated continuously. In many devices, fluctuating current is used due to its own load so alternating current makes it hard to compute the remaining battery level. For many devices, there could be battery level indicator as solution. This solution gives clue about the remaining time for user but it does not give any hint for microcontroller about battery situation. For low cost devices, it could be very difficult to estimate the remaining storage energy in the battery. In this study, microcontroller compatible sealed lead acid battery remaining energy predictor based on adaptive neural fuzzy inference system has been designed and proposed. In order to test proposed method, mean absolute error and leave one out have been used to measure proposed system performance. The obtained mean absolute error results for leave one out is 10.55, epoch error is 11.72. Through the study, low adaptive neural fuzzy inference system rules and low microcontroller memory consumption were aimed.
Web Information Systems and Mining, International Conference on; 11/2009
[Show abstract][Hide abstract] ABSTRACT: In this study, electromyography signals sampled from children undergoing orthodontic treatment were used to estimate the effect of an orthodontic trainer on the anterior temporal muscle. A novel data normalization method, called the correlation- and covariance-supported normalization method (CCSNM), based on correlation and covariance between features in a data set, is proposed to provide predictive guidance to the orthodontic technique. The method was tested in two stages: first, data normalization using the CCSNM; second, prediction of normalized values of anterior temporal muscles using an artificial neural network (ANN) with a Levenberg-Marquardt learning algorithm. The data set consists of electromyography signals from right anterior temporal muscles, recorded from 20 children aged 8-13 years with class II malocclusion. The signals were recorded at the start and end of a 6-month treatment. In order to train and test the ANN, two-fold cross-validation was used. The CCSNM was compared with four normalization methods: minimum-maximum normalization, z score, decimal scaling, and line base normalization. In order to demonstrate the performance of the proposed method, prevalent performance-measuring methods, and the mean square error and mean absolute error as mathematical methods, the statistical relation factor R2 and the average deviation have been examined. The results show that the CCSNM was the best normalization method among other normalization methods for estimating the effect of the trainer.
Proceedings of the Institution of Mechanical Engineers Part H Journal of Engineering in Medicine 11/2009; 223(8):991-1001. DOI:10.1243/09544119JEIM619 · 1.33 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: In this paper, we have proposed a new feature selection method called kernel F-score feature selection (KFFS) used as pre-processing step in the classification of medical datasets. KFFS consists of two phases. In the first phase, input spaces (features) of medical datasets have been transformed to kernel space by means of Linear (Lin) or Radial Basis Function (RBF) kernel functions. By this way, the dimensions of medical datasets have increased to high dimension feature space. In the second phase, the F-score values of medical datasets with high dimensional feature space have been calculated using F-score formula. And then the mean value of calculated F-scores has been computed. If the F-score value of any feature in medical datasets is bigger than this mean value, that feature will be selected. Otherwise, that feature is removed from feature space. Thanks to KFFS method, the irrelevant or redundant features are removed from high dimensional input feature space. The cause of using kernel functions transforms from non-linearly separable medical dataset to a linearly separable feature space. In this study, we have used the heart disease dataset, SPECT (Single Photon Emission Computed Tomography) images dataset, and Escherichia coli Promoter Gene Sequence dataset taken from UCI (University California, Irvine) machine learning database to test the performance of KFFS method. As classification algorithms, Least Square Support Vector Machine (LS-SVM) and Levenberg–Marquardt Artificial Neural Network have been used. As shown in the obtained results, the proposed feature selection method called KFFS is produced very promising results compared to F-score feature selection.
Expert Systems with Applications 09/2009; 36(7-36):10367-10373. DOI:10.1016/j.eswa.2009.01.041 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Due to the fact that there exist only a small number of complex systems in artificial immune systems (AISs) that solve nonlinear problems, there is a need to develop nonlinear AIS approaches that would be among the well-known solution methods. In this study, we developed a kernel-based AIS to compensate for this deficiency by providing a nonlinear structure via transformation of distance calculations in the clonal selection models of classical AIS to kernel space. Applications of the developed system were conducted on Statlog heart disease dataset, which was taken from the University of California, Irvine Machine-Learning Repository, and on Doppler sonograms to diagnose atherosclerosis disease. The system obtained a classification accuracy of 85.93% for the Statlog heart disease dataset, while it achieved a 99.09% classification success for the Doppler dataset. With these results, our system seems to be a potential solution method, and it may be considered as a suitable method for hard nonlinear classification problems.
IEEE transactions on information technology in biomedicine: a publication of the IEEE Engineering in Medicine and Biology Society 05/2009; 13(4):621-8. DOI:10.1109/TITB.2009.2019637 · 2.49 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: In this paper, we have made medical application of a new artificial immune system named the information gain-based artificial immune recognition system (IG-AIRS) which is minimized the negative effects of taking into account all attributes in calculating Euclidean distance in shape-space representation which is used in many artificial immune systems. For medical data, microorganism dataset was applied in the performance analysis of our proposed system. Microorganism dataset was obtained using Cyranose 320 electronic nose. Our proposed system reached 92.35% classification accuracy with five-fold cross validation method. This result ensured that IG-AIRS would be helpful in classification of microorganism species based on laboratory tests, and would open the way to various microorganism species determine support by using electronic nose.
Expert Systems with Applications 04/2009; 36(3-36):5168-5172. DOI:10.1016/j.eswa.2008.06.029 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Generally, many classifier systems compel in the classification of multi-class problems. The aim of this study is to improve the classification accuracy in the case of multi-class classification problems. In this study, we have proposed a novel hybrid classification system based on C4.5 decision tree classifier and one-against-all approach to classify the multi-class problems including dermatology, image segmentation, and lymphography datasets taken from UCI (University of California Irvine) machine learning database. To test the proposed method, we have used the classification accuracy, sensitivity-specificity analysis, and 10-fold cross validation. In this work, firstly C4.5 decision tree has been run for all the classes of dataset used and achieved 84.48%, 88.79%, and 80.11% classification accuracies for dermatology, image segmentation, and lymphography datasets using 10-fold cross validation, respectively. The proposed method based on C4.5 decision tree classifier and one-against-all approach obtained 96.71%, 95.18%, and 87.95% for above datasets, respectively. These results show that the proposed method has produced very promising results in the classification of multi-class problems. This method can be used in many pattern recognition applications. In future, instead of C4.5 decision tree, other classification algorithms such as Bayesian learning, artificial immune system algorithms, artificial neural networks can be used.
Expert Systems with Applications 03/2009; 36(2-36):1587-1592. DOI:10.1016/j.eswa.2007.11.051 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: In this paper, we have made medical application of a new artificial immune system named the information gain based artificial immune recognition system (IG-AIRS) which minimizes the negative effects of taking into account all attributes in calculating Euclidean distance in shape–space representation which is used in many artificial immune systems. For medical data, thyroid disease data set was applied in the performance analysis of our proposed system. Our proposed system reached 95.90% classification accuracy with 10-fold CV method. This result ensured that IG-AIRS would be helpful in diagnosing thyroid function based on laboratory tests, and would open the way to various ill diagnoses support by using the recent clinical examination data, and we are actually in progress.
Expert Systems with Applications 03/2009; 36(2):3086-3092. DOI:10.1016/j.eswa.2008.01.026 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: In this paper, we propose a new feature selection method called class dependency based feature selection for dimensionality reduction of the macular disease dataset from pattern electroretinography (PERG) signals. In order to diagnosis of macular disease, we have used class dependency based feature selection as feature selection process, fuzzy weighted pre-processing as weighted process and decision tree classifier as decision making. The proposed system consists of three parts. First, we have reduced to 9 features number of features of macular disease dataset that has 63 features using class dependency based feature selection, which is first developed by ours. Second, the macular disease dataset that has 9 features is weighted by using fuzzy weighted pre-processing. And finally, decision tree classifier was applied to PERG signals to distinguish between healthy eye and diseased eye (macula diseases). The employed class dependency based feature selection, fuzzy weighted pre-processing and decision tree classifier have reached to 96.22%, 96.27% and 96.30% classification accuracies using 5–10–15-fold cross-validation, respectively. The results confirmed that the medical decision making system based on the class dependency based feature selection, fuzzy weighted pre-processing and decision tree classifier has potential in detecting the macular disease. The stated results show that the proposed method could point out the ability of design of a new intelligent assistance diagnosis system.
Expert Systems with Applications 03/2009; 36(2):2584-2591. DOI:10.1016/j.eswa.2008.02.035 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: The aim of this research was to compare classifier algorithms including the C4.5 decision tree classifier, the least squares support vector machine (LS-SVM) and the artificial immune recognition system (AIRS) for diagnosing macular and optic nerve diseases from pattern electroretinography signals. The pattern electroretinography signals were obtained by electrophysiological testing devices from 106 subjects who were optic nerve and macular disease subjects. In order to show the test performance of the classifier algorithms, the classification accuracy, receiver operating characteristic curves, sensitivity and specificity values, confusion matrix and 10-fold cross-validation have been used. The classification results obtained are 85.9%, 100% and 81.82% for the C4.5 decision tree classifier, the LS-SVM classifier and the AIRS classifier respectively using 10-fold cross-validation. It is shown that the LS-SVM classifier is a robust and effective classifier system for the determination of macular and optic nerve diseases.
[Show abstract][Hide abstract] ABSTRACT: Artificial Immune Recognition System (AIRS) classifier algorithm is robust and effective in medical dataset classification applications such as breast cancer, heart disease, diabetes diagnosis etc. In our previous work, we have proposed a new resource allocation mechanism called fuzzy resource allocation in AIRS algorithm both to improve the classification accuracy and to decrease the computation time in classification process. Here, AIRS and Fuzzy-AIRS classifier algorithms and one against all approach have been combined to increase the classification accuracy of obstructive sleep apnea syndrome (OSAS) that is an important disease that influences both the right and the left cardiac ventricle. The OSAS dataset consists of four classes including of normal (25 subjects), mild OSAS (AHI (Apnea and Hypoapnea Index) = 5-15 and 14 subjects), moderate OSAS (AHI < 15-30 and 18 subjects), and serious OSAS (AHI > 30 and 26 subjects). In the extracting of features that is characterized the OSAS disease, the clinical features obtained from Polysomnography used diagnostic tool for obstructive sleep apnea in patients clinically suspected of suffering from this disease have been used. The used clinical features are Arousals Index (ARI), Apnea and Hypoapnea Index (AHI), SaO2 minimum value in stage of REM, and Percent Sleep Time (PST) in stage of SaO2 intervals bigger than 89%. Even though AIRS and Fuzzy-AIRS classifiers have been used in the classifying multi-class problems, theirs classification performances are low in the case of multi-class classification problems. Therefore, we have used two classes in AIRS and Fuzzy-AIRS classifiers by means of one against all approach instead of four classes comprising the healthy subjects, mild OSAS, moderate OSAS, and serious OSAS. We have applied the AIRS, Fuzzy-AIRS, AIRS with one against all approach (Pairwise AIRS), and Fuzzy-AIRS with one against all approach (Pairwise Fuzzy-AIRS) to OSAS dataset. The obtained classification accuracies are 63.41%, 63.41%, 87.19%, and 84.14% using the above methods for 200 resources, respectively. These results show that the best method for diagnosis of OSAS is the combination of AIRS and one against all approach (Pairwise AIRS).
Journal of Medical Systems 01/2009; 32(6):489-97. DOI:10.1007/s10916-008-9155-7 · 2.21 Impact Factor