International Journal of Intelligent Systems and Applications in Engineering

Published by International Journal of Intelligent Systems and Applications in Engineering
Online ISSN: 2147-6799
Publications
Energy demand estimation of TLBOED between 2006 and 2012 years according to scenario 2.
Energy demand estimation of TLBOED models between 1996 and 2005 years.
Article
In this study, the estimation of Turkey primary electric energy demand until 2035 is tried to estimate by using Teaching-Learning Based Optimization (TLBO) Algorithm. Two models are proposed which are based on economic indicators TLBO algorithm linear energy demand (TLBOEDL) and TLBO algorithm quadratic energy demand (TLBOEDQ). In both of these two models the indicators used are Gross Domestic Product (GDP), population, importation and exportation. After a comparison of these two models with real values between 1979 and 2005 years, it is applied to the estimation of Turkey electric energy demand until 2035 by three different scenario. The estimation results are suitable with the estimation of Turkey total primary energy supply of 2013 Energy Report of World Energy Council Turkish National Committee (WEC-TNC ).
 
Article
FPGA-based embedding system designs have been preferred for industrial applications and prototyping because of the advantages of parallel processing, reconfigurability and low cost. Due to having characteristic structure of the parallel processing of Artificial Neural Networks (ANNs), these systems provide the advantage of speed and performance when they are implemented with FPGA-based hardware. The hardware implementation of transfer functions used for modeling non-linear systems is a challenging problem. Therefore, this problem creates convergence problems. In this paper, non-linear Sprott 94 S system has been modeled using ANNs running on FPGA. All related parameter values and processes are defined with IEEE-754-1985 32-bit floating point number format. ANN-based Sprott 94 S system design has been developed using VHDL synthesized using Xilinx ISE Design Tools. In test stage, ANN-based Sprott 94 S system has been tested using 3X100 data set and obtained error analysis results have been presented. The constructed design has been performed for Xilinx VIRTEX-6 family XC6VHX255T-3FF1923 FPGA chip using Place&Route process and chip usage statistics have been given. The clock frequency of ANN-based Sprott 94 S system which has pipeline processing scheme has been obtained with the value of 304.534 MHz. Accordingly, the proposed FPGA-based ANN system has produced 3X3.284 billion outputs in 1 second.
 
Article
Artificial Neural Networks (ANNs) have emerged as an important tool for classification problem. This paper presents an application of ANN model trained by artificial bee colony (ABC) optimization algorithm for classification the wheat grains into bread and durum. ABC algorithm is used to optimize the weights and biases of three-layer multilayer perceptron (MLP) based ANN. The classification is carried out through data of wheat grains (#200) acquired using image-processing techniques (IPTs). The data set includes five grain’s geometric parameters: length, width, area, perimeter and fullness. The ANN-ABC model input with the geometric parameters are trained through 170 wheat grain data and their accuracies are tested via 30 data. The ANN-ABC model numerically calculate the outputs with mean absolute error (MAE) of 0.0034 and classify the grains with accuracy of 100% for the testing process. The results of ANN-ABC model are compared with other ANN models trained by 4 different learning algorithms. These results point out that the ANN trained by ABC optimization algorithm can be successfully applied to classification of wheat grains.
 
Hourly electrical load data of Turkey  
MAPE errors of hybrid model
Article
Short term load forecasting is a subject about estimating future electricity consumption for a time interval from one hour to one week and it has a vital importance for the operation of a power system and smart grids. This process is mandatory for distribution companies and big electricity consumers, especially in liberalized energy markets. Electricity generation plans are made according to the amount of electricity consumption forecasts. If the forecast is overestimated, it leads to the start-up of too many units supplying an unnecessary level of reserve, therefore the production cost is increased. On the contrary if the forecast is underestimated, it may result in a risky operation and consequently power outages can occur at the power system. In this study, a hybrid method based on the combination of Artificial Bee Colony (ABC) and Artificial Neural Network (ANN) is developed for short term load forecasting. ABC algorithm is used in ANN learning process and it optimizes the neuron connections weights of ANN. Historical load, temperature difference and season are selected as model inputs. While three years hourly data is selected as training data, one year hourly data is selected as testing data. The results show that the application of this hybrid system produce forecast values close to the actual values.
 
Article
While a healthy human walks, his or her legs mutually perform good repeatability with high accuracy. This provides an esthetical movement and balance. People with above knee prosthesis want to perform walking as esthetical as a healthy human. Therefore, to achieve a healthy walking, the above knee prosthesis must provide a good stiffness performance. Especially stiffness values are required when adding a second axis movement to the ankle for eversion and inversion. In this paper, stiffness analysis of above-knee prosthesis is presented. The translational displacement of above knee prosthesis is obtained when the prosthesis is subjected to the external forces. Knowing stiffness values of the above knee prosthesis, designers can compute prosthesis parameters such as ergonomic structure, height, and weight and energy consumption.
 
Speakers demographic data (US: American english accent, Non- US: Non-American accent)
Confusion matrix
Calculation of model evaluation performance criteria
Performance criteria required to evaluate classification success
Article
Sound is the pressure wave created by an object vibrating with a certain frequency. 3 organs are needed for the formation of voice in humans. These are lungs, vocal cords and mouth. Due to the structure of these organs and the similarity of the person with their current language, they can speak another language with different accent. A language can be spoken in different parts of the same country and in different countries. The second most widely used language in the world is English, has numerous accents around the world. In this study, it is aimed to determine which country the English accent spoken in different regions belongs to. In the dataset used, there are 330 sound samples including English accents spoken in Spain, France, Germany, Italy, England and America. Classification has been made with 12 features obtained by Mel Frequency Cepstrum Coefficients feature extraction method. k-Nearest Neighbor (kNN) were used in the classification and 87.2% success was achieved.
 
Ratings and Parameters of IM 
Article
This paper presents a design of a fuzzy logic controller (FLC) with tuning output scaling factor for speed control of indirect field oriented induction motor (IM) taking core loss into account. The variation of output scaling factor of FLC depends on the normalized output of FLC. Firstly the speed control of IM taking core loss into account is presented by using FLC with fixed scaling factors (FLC-FSF). Secondly the speed controller based on suggested FLC with tuning output scaling factor (FLC-TOSF) is proposed. The performance of the proposed FLC-TOSF for speed control of IM are investigated and compared to those obtained using FLC-FSF at different operating conditions and variation of parameters. A comparison of simulation results shows that the convergence of actual speed to reference speed is faster by using the proposed FLC-TOSF.
 
Article
Abstract: Compared to traditional motor controlling techniques, modern AI controllers have many advantages. However, most of the developed FOC mechanisms are based on classical controlling techniques such as PID controllers, hybrid AI-classical controllers and model reference controllers. All these traditional controllers are based on sophisticated mathematical models (system transfer functions). These traditional controllers are unable to tune its system parameters by it-self to adapt according to the non-linear variations of actual speed and torque of the motor. This paper discussed, a novel scheme of metaheuristic adaptive fuzzy logic-particle swarm optimization control mechanism to optimize the speed regulation of electric current space vector-controlled BLDC motor. Therefore, dynamic TSK-PSO-FLC was investigated. The dynamic behaviour of the proposed controller enables it to optimize its tuning parameters by it-self under non-linear load and speed varying conditions to track the desired angular speed and the torque trajectories. This is a part of the designed and developed dynamic AI controller, for stability and traction control of an all-wheel-drive electric rover. Therefore, initially, the performance of the proposed controller has been tested on a simulation environment (MATLAB Simulink model) for one wheel. Finally, the identified dynamic parameters through the Simulink model (the sensored BLDC motor and the proposed AI controller) were utilized to test the performance of the developed TSK-PSO-FLC, while it is tracking a given desired speed trajectory of the BLDC motor (sensored, 3-ph and 250 W) in real-time operation. Simulated test results are analyzed and compared with the observed test results through the developed hardware model in addition to the newly published research work. The angular speed of 2500 rpm within a 500 N m torque have been taken into consideration as a generalized condition. It was noticed that the percentage overshoot (Mp%), settling time (Ts) and the steady-state error (Ess) is 0.501, 731.455 μs and 1.22 respectively. Therefore, compared to classical control based FOC mechanisms, the analyzed test results of the proposed control mechanism is showing that it has been optimized and enhanced the speed regulation performance of the BLDC motor significantly while increasing the frequency of the desired input trajectory up to 2 kHz. Keywords: Brushless direct current (BLDC) motor, Fuzzy logic (FL), Field-oriented control (FOC), Particle swarm optimization (PSO)
 
Summarization of Activity Datasets
Article
Human action recognition is an important area of research in the field of computer vision due to its extensive applications like security surveillance; content based video retrieval and annotation, human computer interaction, human fall detection, video summarization, robotics, etc. The surveillance system deals with the monitoring and analysing the human behaviour and activities. The main aim of the smart surveillance system is to recognize anomalous behaviour in given scene and provide real time intimation to relevant person. We have designed and tested Smart Surveillance System for College Corridor Scene (3S2CS). The system recognises the anomalous behaviour and an intimation is provided in the form of Firebase Cloud Messaging (FCM) alert on the android mobile phone to the authorised user. This paper mainly discuss the methodologies used for the human action recognition. The basic step is to provide video as an input. These videos are further divided into number of frames. The videos are used for training and for each video, Scale Invariant Feature Transform (SIFT) is applied for extracting features and developing feature vectors. The actions are classified using K Nearest Neighbour (KNN)and Support Vector Machine (SVM) classifier. Two standard offline datasets considered for testing are Weizmann and UTD-MHAD. For real time scenario we have created dataset in college campus called as College Corridor dataset. It contains student activities like falling, fighting, walking, running, sitting and other general actions. If falling or fighting action is detected, thenotification is sent to the authorized user who has installed ActionDetector android application and registered a device for the same. Action recognition accuracy is 92.91% using SVM and 90.83% using KNN (PDF) Human Action Recognition on Real Time and Offline Data. Available from: https://www.researchgate.net/publication/332040006_Human_Action_Recognition_on_Real_Time_and_Offline_Data [accessed Mar 30 2019].
 
Article
The utilization of electrical power system has been rising frequently from past to now and there is a need of dependable electrical transmission and distribution networks so as to ensure continuous and balanced energy. Besides, conventional energy governance systems have been forced to change as a result of rises in the usage of renewable energy resources and the efficiency of demand-side on the market. In this regard, electrical power systems should be planned and operated, appropriately and the balance of production and consumption demand should be provided within the nominal voltage limits. In this study, firstly, the current status of Marmara region interconnected power grid in Turkey is evaluated. Afterwards, the multiple cascading failure outages scenarios are modeled by “DIgSILENT Power Factory V14” software. The critical transmission line scenarios are implemented on the high voltage power grid model improved. These scenarios are based on the period of maximum and minimum production and consumption demand and the effects of demand response in this period. As a result of grid vulnerability analyses performed, several findings has been obtained about the impacts of different line scenarios on the high voltage transmission system, the optimization of power grid voltage profile and the role of production and consumption demand response on voltage regulation. The utilization of electrical power system has been rising frequently from past to now and there is a need of dependable electrical transmission and distribution networks so as to ensure continuous and balanced energy. Besides, conventional energy governance systems have been forced to change as a result of rises in the usage of renewable energy resources and the efficiency of demand-side on the market. In this regard, electrical power systems should be planned and operated, appropriately and the balance of production and consumption demand should be provided within the nominal voltage limits. In this study, firstly, the current status of Marmara region interconnected power grid in Turkey is evaluated. Afterwards, the multiple cascading failure outages scenarios are modeled by “DIgSILENT Power Factory V14” software. The critical transmission line scenarios are implemented on the high voltage power grid model improved. These scenarios are based on the period of maximum and minimum production and consumption demand and the effects of demand response in this period. As a result of grid vulnerability analyses performed, several findings has been obtained about the impacts of different line scenarios on the high voltage transmission system, the optimization of power grid voltage profile and the role of production and consumption demand response on voltage regulation.
 
Article
This paper addresses the issue of controller design for a class of multi-input multi-output (MIMO) uncertain underactuated systems with saturating inputs. A systematic controller framework, composed of a hierarchically generated control term, meant to ensure the stabilization of a particular portion of system dynamics and some dedicated control terms designed to solve the tracking problem of the remaining system dynamics is presented. Wavelet neural networks are used as adaptive tuners to approximate the system uncertainties also to reshape the control terms so as to deal with the saturation nonlinearity in an antiwindup paradigm. Gradient based tuning laws are developed for the online tuning of adjustable parameters of the wavelet network. A Lyapunov based stability analysis is carried out to ensure the uniformly ultimately bounded (UUB) stability of the closed loop system. Finally, a simulation is carried out which supports the theoretical development.
 
Article
Power converters are generally utilized to convert the power from the wind sources to match the load demand and grid requirement to improve the dynamic and steady-state characteristics of wind generation systems and to integrate the energy storage system to solve the challenge of the discontinuous character of the renewable energy. In the low-voltage wind energy systems, interleaved boost converters (IBC) are often used to operate high currents in the system. IBCs are extremely sensitive to the constantly changing loading conditions. These situations require a robust control operation which can ensure a sufficient performance of the IBC over a large-scale changing load. Neural networks (NN) have emerged over the years and have found applications in many engineering fields, including control. In this paper, the adaptive control of interleaved boost converter with power factor correction (PFC) is investigated for grid-connected synchronous generator of wind energy system. For this purpose, a model reference adaptive control (MRAC) based on NN is proposed. Analysis results show that the proposed control strategy for the IBCs achieves near unity power factor (PF) and low total harmonic distortion (THD) in a wide operating range.
 
Article
Mapping human cognition into automated analysis is the key area of research due to its fascinating applications in almost every area of developing artificially intelligent machines. The best way to understand the functioning of brain is to study electroencephalogram (EEG) patterns, therefore, a lot of research has been directed towards studying EEG signals. Since, EEG recordings are subject dependent and exhibit variations due to external influences or type of recording instruments, it is hard to develop a generalized affect categorization system that can provide robust affect labelling to the EEG patterns. To overcome this, proposed work presents a novel general framework for affect-based cognitive analysis. The proposed system involves following steps: pre-processing, feature selection, Generalized Procrustes Analysis (GPA) step to reduce the inter-class and intra-class variance and finally, the processed pattern is passed on to a trained classifier for classifying the pattern into appropriate affect categories. The presented approach has been tested on single as well as multiple subjects’ EEG data taken from two different datasets, Database for Emotion Analysis using Physiological signals (DEAP) and SJTU Emotion EEG Dataset (SEED) and performance of popular classifiers are assessed. The experimental results suggest that SVM classifier is the best among the selected ones for classifying single as well as mixed subjects’ data.
 
Article
In this study, we have provided an alternative solution to spam and legitimate email classification problem. The different deep learning architectures are applied on two feature selection methods, including the Mutual Information (MI) and Weighted Mutual Information (WMI). Firstly, feature selection methods including WMI and MI are applied to reduce number of selected terms. Secondly, the feature vectors are contructed with concept of bag-of-words (BoW) model. Finally, the performance of system is analysed with using Artificial Neural Network (ANN), Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BILSTM) models. After experimental simulations, we have observed that there is a competition between detection results of using WMI and MI when commented with accuracy rates for the agglutinative language, namely Turkish. The experimental scores shows that the LSTM and BILSTM gives 100% accuracy scores when combined with MI or WMI, for spam and legitimate emails. However, for particular cross validation, the performance WMI is higher than MI features in terms e-mail grouping. It turns out that WMI and MI with deep learning architectures seems more robust to spam email detection when considering the high detection scores.
 
Article
Air is one of the most important life sources for all living things. Gases that are present and absent in the composition of clean air also consideredas pollutants in the atmosphere. If the pollutants rise above a certain concentration level, air pollution occurs. Air pollution damages all living things, especially human health. Accurate estimation of pollutant concentrations through air pollution modeling has an important effect in reducing the adverse effects of air pollution and taking necessary precautions. Conventional statistical models are widely used in air pollution forecasting and modeling. As a different approach, in this study, fuzzy logic algorithm(FLA), which has been increasingly successful in many field applications, isused to model air quality and air pollution analyzes aremade based on this model. Ankara province data is used in the sample of the research .
 
Article
In this work, we will adapt the NAW (Nagri, Adler and Wetzel) precipitation, estimation approach to the north Algeria events using the Meteosat Second Generation (MSG) satellite images. The tests are carried out on seven areas of northern Algeria: Sidi Bel Abbes, Oran Port, Algiers Port, Dar El Beida, Bedjaia, Jijel-Achouat and Annaba, in winter 2006. The NAW approach is applied by thresholding to temperature from 253 K. The validation is performed by comparaison the estimated rainfall to in situ measures collected by the National Office of Meteorology in Dar El Beida (Algeria). We use the infrared data (10.8µm channel) of SEVIRI sensor in this study. The results obtained indicate that the NAW approach gives satisfactory results for the rain rates: 4mm/h assigned to the coldest 10%, 2mm/h assigned to the next 40% and 0mm/h given to the remaining 50% of the area defined as cloud. The rain rate 8mm/h assigned to the coldest 10% of the pixels in the cloud applied for the convective clouds observed for tropical regions are not valid for the Algerian climate, especially for the stratiform clouds type.
 
Article
Graph coloring problem (GCP) is getting more popular to solve the problem of coloring the adjacent regions in a map with minimum different number of colors. It is used to solve a variety of real-world problems like map coloring, timetabling and scheduling. Graph coloring is associated with two types of coloring as vertex and edge coloring. The goal of the both types of coloring is to color the whole graph without conflicts. Therefore, adjacent vertices or adjacent edges must be colored with different colors. The number of the least possible colors to be used for GCP is called chromatic number. As the number of vertices or edges in a graph increases, the complexity of the problem also increases. Because of this, each algorithm can not find the chromatic number of the problems and may also be different in their executing times. Due to these constructions, GCP is known an NP-hard problem. Various heuristic and metaheuristic methods have been developed in order to solve the GCP. In this study, we described First Fit (FF), Largest Degree Ordering (LDO), Welsh and Powell (WP), Incidence Degree Ordering (IDO), Degree of Saturation (DSATUR) and Recursive Largest First (RLF) algorithms which have been proposed in the literature for the vertex coloring problem and these algorithms were tested on benchmark graphs provided by DIMACS. The performances of the algorithms were compared as their solution qualities and executing times. Experimental results show that while RLF and DSATUR algorithms are sufficient for the GCP, FF algorithm is generally deficient. WP algorithm finds out the best solution in the shortest time on Register Allocation, CAR, Mycielski, Stanford Miles, Book and Game graphs. On the other hand, RLF algorithm is quite better than the other algorithms on Leighton, Flat, Random (DSJC) and Stanford Queen graphs.
 
Article
Breast cancer is one of the most common types of cancer and is the second main cause of cancer death in females. Early detection of breast cancer is crucial for the survival of a patient as well as for the quality of life throughout cancer treatment. The aim of this study is to develop improved machine learning models for early diagnosis of breast cancer with high accuracy. In this context, a performance comparison of machine learning algorithms including Support Vector Machines, Decision Trees, Naive Bayes, K-Nearest Neighbor, and Ensemble Classifiers was performed on a dataset consisting of routine blood analysis combined with anthropometric measurements to diagnose breast cancer. Neighborhood component analysis was applied as a feature selection method to reveal relevant biomarkers that can be used in breast cancer prediction. In order to assess the performance of each proposed classifier model, two different data division procedures such as hold-out and 10-fold cross-validation were employed. Bayesian Optimization algorithm was applied to all classifiers for the maximizing the prediction accuracy. Different performance criteria such as accuracy, precision, sensitivity, specificity, and F-measure were used to measure the success of each classifier. Experimental results show that the Bayesian optimization-based K-Nearest Neighbor performs better than other machine learning algorithms under the hold-out data division protocol with an accuracy of 95.833%. The results obtained in this study may provide a new perspective on the application of improved machine learning techniques for the early detection of breast cancer.
 
Article
Target detection in hyperspectral images is important in many applications including search and rescue operations, defence systems, mineral exploration and border security. For this purpose, several target detection algorithms have been proposed over the years, however, it is not clear which of these algorithms perform best on real data and on sub-pixel targets, and moreover, which of these algorithms have complementary information and should be fused together. The goal of this study is to detect the nine arbitrarily placed sub-pixel targets, from seven different materials from a 1.4km altitude. For this purpose, eight signature-based hyperspectral target detection algorithms, namely the GLRT, ACE, SACE, CEM, MF, AMSD, OSP and HUD, and three anomaly detectors, namely RX, Maxmin and Diffdet, were tested and compared. Among the signature-based target detectors, the three best performing algorithms that have complementary information were identified. Finally these algorithms were fused together using four different fusion algorithms. Our results indicate that with a proper fusion strategy, five of the nine targets could be found with no false alarms.
 
Percentage average accuracy of statistical and artificial intelligence based classification algorithms on the experimental data  
Article
A variety of methods are used in order to classify cancer gene expression profiles based on microarray data. Especially, statistical methods such as Support Vector Machines (SVM), Decision Trees (DT) and Bayes are widely preferred to classify on microarray cancer data. However, the statistical methods can often be inadequate to solve problems which are based on particularly large-scale data such as DNA microarray data. Therefore, artificial intelligence-based methods have been used to classify on microarray data lately. We are interested in classifying microarray cancer gene expression by using both artificial intelligence based methods and statistical methods. In this study, Multi-Layer Perceptron (MLP), Radial basis Function Network (RBFNetwork) and Ant Colony Optimization Algorithm (ACO) have been used including statistical methods. The performances of these classification methods have been tested with validation methods such as v-fold validation. To reduce dimension of DNA microarray gene expression has been used Correlation-based Feature Selection (CFS) technique. According to the results obtained from experimental study, artificial intelligence-based classification methods exhibit better results than the statistical methods.
 
Article
There are various applications using computer-aided quality controlling system. In this study, seed data set acquired from UCI machine learning database was used. The purpose of the study is to perform the operations for separation of seed species from each other in the seed data set. Three different seed whose data was acquired from the UCI machine learning database was used. Later it was classified by applying the methods of KNN, Naive Bayes, J48 and multilayer perceptron to the dataset. While wheat seed data received from the UCI machine learning database was classified, WEKA program was used. Depending on the number of neurons the highest classification success came in 7-layer neurons. Our success rate for the number of 7-layer neurons came to 97.17% When the classification success rate was calculated according to KNN for the values of different neighbour, the highest success rate for neighbour was set at 95.71% for 4. Neighbour. With this method, classification of seeds depending on their properties was provided more quickly and effectively.
 
Generative model for LDA.
Extracted multi-words and their types
Graphical model for LDA.
One of the examples of extracted topics
Ranked lists of aspect-sentiment pairs
Article
Online user reviews have a great influence on decision-making process of customers and product sales of companies. However, it is very difficult to obtain user sentiments among huge volume of data on the web consequently; sentiment analysis has gained great importance in terms of analyzing data automatically. On the other hand, sentiment analysis divides itself into branches and can be performed better with aspect level analysis. In this paper, we proposed to extract aspect-sentiment pairs from a Turkish reviews dataset. The proposed task is the fundamental and indeed the critical step of the aspect level sentiment analysis. While extracting aspect-sentiment pairs, an unsupervised topic model Latent Dirichlet Allocation (LDA) is used. With LDA, aspect-sentiment pairs from user reviews are extracted with 0.86 average precision based on ranked list. The aspect-sentiment pair extraction problem is first time realized with LDA on a real world Turkish user reviews dataset. The experimental results show that LDA is effective and robust in aspect-sentiment pair extraction from user reviews.
 
Article
India is the world's second-largest cement manufacturer. It is a significant contributor to the Indian economy's GDP. Kadapa district is one of Andhra Pradesh's largest cement producers. Limestone and cement plant sediment and mine have local, regional, and global impacts on soil, vegetation, and water and air quality. As a result, mapping and change evaluation of the mining area are critical for the sustainability of the cement industry. With today's advancements in remote sensing technology, mapping the Earth's characteristics, observing environmental changes, and controlling natural resources become more efficient with less human efforts than conventional methods. Proposed work focused on land environment temporal change assessment in YSR kadapa district, Andhra Pradesh, India over a period from 1991 to 2019. The results of Landsat-5/7/8 image Hybrid classification using ERDAS IMAGINE and ArcGIS over the study area 684KM2 showing an overall accuracy 92 % and kappa index 0.9 in comparison to conventional methods of classification.
 
Top-cited authors
Murat Koklu
  • Selcuk University
Kadir Sabanci
  • Karamanoglu Mehmetbey Üniversitesi
Abdullah Erdal Tümer
  • Necmettin Erbakan Üniversitesi
Ilker Ali Ozkan
  • Selcuk University
Muhammet Fatih Aslan
  • Karamanoglu Mehmetbey Üniversitesi