Showing papers in "International Journal of Intelligent Engineering and Systems in 2020"
TL;DR: The results of simulation and comparison indicate the superiority and optimal quality of the proposed DGO algorithm over the mentioned algorithms.
Abstract: In this paper, a novel game-based optimization technique entitled darts game optimizer (DGO) is proposed. The novelty of this investigation is DGO designing based on simulating the rules of Darts game. The key idea in DGO is to get the most possible points by the players in their throws towards the game board. Simplicity of equations and lack of control parameters are the main features of the proposed algorithm. The ability and quality of DGO performance in optimization is evaluated on twenty-three objective functions, and then is compared with eight other optimization algorithms including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Gravitational Search Algorithm (GSA), Teaching Learning-Based Optimization (TLBO), Grey Wolf Optimizer (GWO), Grasshopper Optimization Algorithm (GOA), Whale Optimization Algorithm (WOA), and Marine Predators Algorithm (MPA). The results of simulation and comparison indicate the superiority and optimal quality of the proposed DGO algorithm over the mentioned algorithms.
83 citations
TL;DR: The results and data obtained from applying FGBO and other mentioned algorithms on unimodal test functions, multimodalTest functions, and energy commitment problem show that F GBO is able to provide better results in comparison with other well-known optimization algorithms.
Abstract: Heuristic optimization algorithms are widely used to solve problems in different fields of science. In this paper, a new game based optimization method called football game based optimization (FGBO) is presented which simulates the game of football. The population of FGBO are clubs and the variables of the problem are the players belonging to the clubs. FGBO has four phases: a) league holding, b) player transfer, c) practice, and d) promotion and relegation. The power of FGBO in solving optimization problems has been investigated on several benchmark test functions. The result of FGBO and other algorithm are obtained from implantation of these algorithms on unimodal, multimodal, and fixed-dimension multimodal benchmark test functions. Eight optimization algorithms called Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Teaching Learning Based Optimization (TLBO), Grey Wolf Optimizer (GWO), Grasshopper Optimization Algorithm (GOA), Emperor Penguin Optimizer (EPO), Shell Game Optimization (SGO), and Hide Objects Game Optimization (HOGO) have been used to compare these results. The proposed FGBO algorithm is also used to solve the energy commitment (EC) problem. Based on the simulation studies and obtained results, FGBO has a higher efficiency than a number of other algorithms. The results and data obtained from applying FGBO and other mentioned algorithms on unimodal test functions, multimodal test functions, and energy commitment problem show that FGBO is able to provide better results in comparison with other well-known optimization algorithms.
50 citations
TL;DR: It has been determined that Multi-Leader optimizer (MLO) has a higher ability to solve optimization problems than existing optimization algorithms.
Abstract: Optimization is a topic that has always been discussed in all different fields of science. One of the most effective techniques for solving such problems is optimization algorithms. In this paper, a new optimizer called Multi-Leader optimizer (MLO) is developed in which multiple leaders guide members of the population towards the optimal answer. MLO is mathematically modelled based on the process of advancing members of the population and following the leaders. MLO performance in optimization is examined on twenty-three standard objective functions. The results of this optimization are compared with the results of the other eight existing optimization algorithms including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Teaching-Learning-Based Optimization (TLBO), Gray Wolf Optimizer (GWO), Grasshopper Optimization Algorithm (GOA), Emperor Penguin Optimizer (EPO), Shell Game Optimization (SGO), and Hide Objects Game Optimization (HOGO). Based on the analysis of the simulation results on unimodal test functions to evaluate exploitation ability and multimodal test functions in order to evaluate exploration ability, it has been determined that MLO has a higher ability to solve optimization problems than existing optimization algorithms.
48 citations
36 citations
TL;DR: A classification model based on supervised machine learning techniques and word-based N-gram analysis to classify Twitter messages automatically into credible and not credible and experiments show that the proposed model achieved an improvement when compared to two models existing in the literature.
Abstract: With the evolution of social media platforms, the Internet is used as a source for obtaining news about current events. Recently, Twitter has become one of the most popular social media platforms that allows public users to share the news. The platform is growing rapidly especially among young people who may be influenced by the information from anonymous sources. Therefore, predicting the credibility of news in Twitter becomes a necessity especially in the case of emergencies. This paper introduces a classification model based on supervised machine learning techniques and word-based N-gram analysis to classify Twitter messages automatically into credible and not credible. Five different supervised classification techniques are applied and compared namely: Linear Support Vector Machines (LSVM), Logistic Regression (LR), Random Forests (RF), Naïve Bayes (NB) and K-Nearest Neighbors (KNN). The research investigates two feature representations (TF and TF-IDF) and different word N-gram ranges. For model training and testing, 10-fold cross validation is performed on two datasets in different languages (English and Arabic). The best performance is achieved using a combination of both unigrams and bigrams, LSVM as a classifier and TF-IDF as a feature extraction technique. The proposed model achieves 84.9% Accuracy, 86.6% Precision, 91.9% Recall, and 89% F-Measure on the English dataset. Regarding the Arabic dataset, the model achieves 73.2% Accuracy, 76.4% Precision, 80.7% Recall, and 78.5% F-Measure. The obtained results indicate that word N-gram features are more relevant for the credibility prediction compared with content and source-based features, also compared with character N-gram features. Experiments also show that the proposed model achieved an improvement when compared to two models existing in the literature.
29 citations
TL;DR: In this article, the adaptive mask region-based convolutional network (mask-RCNN) was utilized for multi-class object detection in remote sensing images, where transfer learning, data augmentation and fine-tuning were adopted to overcome objects scale variability, small size, the density of objects, and the scarcity of annotated remote sensing image.
Abstract: Fast and automatic object detection in remote sensing images is a critical and challenging task for civilian and military applications. Recently, deep learning approaches were introduced to overcome the limitation of traditional object detection methods. In this paper, adaptive mask Region-based Convolutional Network (mask-RCNN) is utilized for multi-class object detection in remote sensing images. Transfer learning, data augmentation, and fine-tuning were adopted to overcome objects scale variability, small size, the density of objects, and the scarcity of annotated remote sensing image. Also, five optimization methods were investigated namely: Adaptive Moment Estimation (Adam), stochastic gradient decent (SGD), adaptive learning rate method (Adelta), Root Mean Square Propagation (RMSprop) and hybrid optimization. In hybrid optimization, the training process begins Adam then switches to SGD when appropriate and vice versa. Also, the behaviour of adaptive mask RCNN was compared to baseline deep object detection methods. Several experiments were conducted on the challenging NWPU-VHR-10 dataset. The hybrid method Adam_SGD acheived the highest Accuracy precision, with 95%. Experimental results showed detection performance in terms of accuracy and intersection over union (IOU) boost of performance up to 6%.
29 citations
TL;DR: The Rough Set Theory (RST) technique is used to select the most relevant features, which helps to provide the efficient classification of medical data and disease detection, which showed that the RST-RNN method achieved accuracy of 98.57%, where the existing Support Vector Machine (SVM) achieved 90.57% accuracy.
Abstract: In a modern life, early healthcare prediction plays an important role to prevent the loss of life caused by prediction delays in treatment. Nowadays, the researchers focused on the Big data analysis, which is used to identify the future health status and provides an efficient way to overcome the issues in early prediction. Many researches are going on predictive analytics using machine learning techniques to provide a better decision making. Big data analysis provides great opportunities to predict future health status from health parameters and provide best outcomes. However, the data classification is one of the major challenging tasks due to noisy data or missing data in the dataset. Feature selection techniques play an important role in the classification process by removing irrelevant features from the extracted data. In this research work, the Rough Set Theory (RST) technique is used to select the most relevant features, which helps to provide the efficient classification of medical data and disease detection. The selected features are given as input to the Recurrent Neural Network (RNN) technique for disease prediction. The proposed method is also called as RST-RNN, where the experiments are carried out on the UCI machine learning repository dataset in terms of accuracy, f-measure, sensitivity and specificity. The results showed that the RST-RNN method achieved accuracy of 98.57%, where the existing Support Vector Machine (SVM) achieved 90.57% accuracy and Naive Bayes (NB) achieved 97.36% accuracy for heart disease dataset.
28 citations
TL;DR: This paper presents a meta-modelling framework for estimating the energy requirements of microgrids and its applications, and some examples from around the world show the need for more concerted efforts to develop smart grids.
Abstract: Department of Electrical and Electronics Engineering, Shiraz University of Technology, Shiraz, Iran Department of Electrical Engineering, University of Calgary, Calgary Alberta, Canada Department of Electrical Engineering, Faculty of Engineering, University of Shahreza, Shahreza 86481-41143, Iran Center for Research on Microgrids (CROM), Department of Energy Technology, Aalborg University, Aalborg, Denmark * Corresponding author’s Email: adanbax@gmail.com
27 citations
TL;DR: This paper proposes the data-sharing system scheme, which uses a blockchain-based decentralized network that each node can be connected directly to each other, to support the exchange of data between them.
Abstract: One thing that tourists need to plan their tourism activities is a recommendation system. The tourism destinations recommendation system in this study has three primary nodes, namely user, server, and sensor. Each node requires the ability to share data to produce recommendations that the user expects through their mobile devices. In this paper, we propose the data-sharing system scheme uses a blockchain-based decentralized network that each node can be connected directly to each other, to support the exchange of data between them. The block architecture used in the blockchain network has three main parts, namely block information, hashes, and data. Each type of node has a different structure and direction of data communication. Where the user node sends destination assessment data to the server node, then the server node sends data from the machine learning process to the user node. The sensor sends dynamic data about popularity, traffic, and weather to the user node as consideration for finalizing the generating recommendations process. In the process of sending data, each node in the blockchain network goes through several functions, including hashing, block validation, chaining block, and broadcast. We conduct web-based experiments and analysis of the data-sharing system to illustrate the system works. The experimental results show that the system handles data circulation with an average time of mine is 84.5 ms in sending multi-criteria assessment data from the user and 119.1 ms in sending data of machine learning result from the server.
18 citations
TL;DR: The performance of the proposed IOT-ANN-SPTS system is compared with the existing previous related works and it is shown that the proposed system produced less error when compared to the other methods.
Abstract: Advancement of the public transport system is important to modern society for reliable performance. Intelligent public transport system can utilize the time very effectively to give better performance to the society. Fast advancement in equipment, programming, and correspondence innovations has encouraged the rise of Internetassociated devices that give perceptions and information gathering from modern reality. By interfacing, an internetenabled device with the public transport system leads to the intelligent public transport system. This paper proposed intelligent public transport with IoT enabled system to give an accurate prediction to the arrival time of the bus to the particular bus stops. Here, Artificial Neural Network (ANN) is used as a prediction algorithm and ANN is trained with different traffic parameters and environmental conditions. Parameters which are considered in the proposed system includes Distance(D), Waiting Time at Stops (WTS), Red signal Duration at Traffic Signal (RSD), Traffic Density (TD), Turning Density (TRD), Rush hours (RH), Weekends (WE), Weather conditions (WC), Number of passengers in the bus (NP), Public Holiday (PH), Road Type (RT). These parameters and the measured real-time arrival time of the bus in different stops for 10 days is used for training the ANN. This trained ANN is implemented on the server-side. In this paper, the performance of the proposed IOT-ANN-SPTS system is compared with the existing previous related works. From the performance analysis, it is shown that the proposed system produced less error (MAE=134.582, RMSE=197.738, MAPE=28.87) when compared to the other methods.
16 citations
TL;DR: An automated Computer Aided Diagnostic (CAD) system based on chest X-Ray image analysis indicates that the suggested CAD scheme can be a promising supplementary COVID-19 diagnostic tool for clinical doctors.
Abstract: COVID-19 is a vital zoonotic illness caused by Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2) COVID-19 is a very wide-spread among humans thus the early detection and curing of the disease offers a high opportunity of survival for patients Computed Tomography (CT) plays an important role in the diagnosis of COVID-19 As chest radiography can give an indicator of coronavirus Though, an automated Computer Aided Diagnostic (CAD) system for COVID-19 based on chest X-Ray image analysis is presented in this article It is designed for COVID-19 recognition from other MERS, SARS, and ARDS viral pneumonia The optimal threshold value for the segmentation of a chest image is deduced by exploiting Li s' method and particle swarm intelligence Laws' masks are then applied to the segmented chest image for secondary characteristics highlighting After that, nine different vectors of attributes are extracted from the Grey Level Co-occurrence Matrix (GLCM) representation of each Law's mask result Support vector machine ensemble models are then built based on the extracted feature vectors Finally, a weighted voting method is utilized to combine the decisions of ensemble classifiers Experimental findings show an accuracy of 98 04 % It indicates that the suggested CAD scheme can be a promising supplementary COVID-19 diagnostic tool for clinical doctors © 2020, Intelligent Network and Systems Society
TL;DR: The experimental results show that the improved method of parallel model detection, designed in two phases, successfully recognizes and differentiates XOR, AND and OR patterns correctly in all scenarios and sounds in all discovered model and gets 100% fitness.
Abstract: The existing method of graph-based process model discovery has weaknesses in detecting parallel relationship (XOR, AND, and OR). The algorithm only works on a particular graph structure, so it must be reconfigured when applied to other different structures. To answer this problem, this paper proposes an improved method of parallel model detection, which is designed in two phases. The first one consists of three steps; firstly is to count and record the value of relationship frequency into every node in a graph model. Then, the second step implements the algorithm to discover the concurrent relationship. The third step detects all possible split and join relationships. Based on the first phase, then a consistent and robust parallel discovery algorithm can be developed. The first parallel algorithm is to identify the XOR relationship. This algorithm is designed with the rule that the XOR pattern cannot have a concurrent relationship between its branch nodes. Next, the algorithm for detecting AND and OR must detect the existence of any concurrent relationship in its branches. Then, AND and OR pattern is differentiated by their unique characteristic of relationship frequency at branch nodes. To verify the ability of the proposed methods in which the existing method fails, we have designed four scenarios. Scenario 1 and 2 consecutively were arranged with two and three branches parallel model. Scenario 3 located the AND and OR inside the XOR pattern. In scenario 4 the sequence relationships were inserted between split and join of parallel patterns. The experimental results show that the proposed method successfully recognizes and differentiates XOR, AND and OR patterns correctly in all scenarios. It also sounds in all discovered model and get 100% fitness.
TL;DR: This study introduces a new approach that combines naïve Bayes classification (NBC) and genetic algorithm (GA) optimization procedures to effectively explore the search space based on a sample of experimental points and shows that the proposed method could significantly increase the yield compared to the other imputation methods and without feature selection.
Abstract: In the case of high dimensional data with missing values, the process of collecting data from various sources may be miss accidentally, which affected the quality of learning outcomes. a large number of machine learning methods can be applied to explore the search area for imputation and selection of features and parameters. ML classification needs preprocessing with self-organizing map imputation (SOMI) before the imputation of missing values is done to improve the accuracy of the model. This study introduces a new approach that combines naïve Bayes classification (NBC) and genetic algorithm (GA) optimization procedures to effectively explore the search space based on a sample of experimental points. GA is a classification model approach based on the selection of features that cause computational problems, such as reduced dimensions, uncertainty and imbalanced data sets with various classes. In the experiment, preprocessing the data using SOMI yielded error results that were up to 10% for various data sets with missing data compared to other methods. In the SOMI-GANB hybrid model, the experimental results show that the proposed method can significantly improve accuracy by up to 90% compared to other imputation methods and without feature selection. SOMI can be used for homogeneous, heterogeneous and mixed data sets. The results from the experiment clearly showed that the proposed method could significantly increase the yield compared to the other imputation methods and without feature selection. The combination of GA and naïve Bayes classification was chosen because they are simple, easy-to-understand methods that are very effective in finding optimal solutions from a set of possible solutions. Naïve Bayes imputation had higher accuracy compared to neural network imputation.
TL;DR: The most relevant feature subset generated by GRFE in the previous work is employed to assess and compare the performance of a single machine learning technique (Lazy IBK) over an ensemble technique (Random Committee) while detecting intrusions in a computer network.
Abstract: Security has been a crucial factor in this modern digital period due to the rapid development of information technology, which is followed by serious computer crimes that, in turn, led to the emergence of Intrusion Detection Systems (IDSs). Various approaches such as single machine learning classifiers and Ensemble Classifiers couple with features selection methods have been proposed to improve the performance of IDS. In this regard, in the previous work, we have used the NSL-KDD IDS dataset, Gain Ratio Feature Evaluator (GRFE), and Correlation Ranking Filter (CRF) feature selection methods coupled with various machine-learning techniques to detect intrusions in computer network traffic. While the experiment has demonstrated that GRFE selects the most relevant feature subsects over CRF, which results in different performance, the previous work can be extended as follows. First, the most relevant feature subset generated by GRFE in the previous work is employed to assess and compare the performance of a single machine learning technique (Lazy IBK, aka K-Nearest Neighbor) over an ensemble technique (Random Committee) while detecting intrusions in a computer network. Second, two distinct datasets (NSL-KDD and UNSW-NB15) are employed for better performance analysis. Third, limitations encountered in the domain of network intrusion detection are also discussed. The results reveal that the ensemble technique performs well over a single machine learning technique with a misclassification gap of 0.969% and 1.19% (obtained using NSL-KDD dataset) and 1.62% and 1.576% (obtained using UNSW-NB15 dataset).
TL;DR: This research proposed an algorithm for cell separation based on enhanced edge detection and edge linking for automatic identification of L1, L2, and L3, in touching cell in Acute Lymphoblastic Leukemia.
Abstract: Acute lymphoblastic leukemia (ALL) is a type of blood cancer that begins with immature lymphocytes in bone marrow. The challenge in automatic identification of ALL when there are touching cells in image. The previous studies related to the separation of touching cells are still constrained by oversegmentation, undersegmentation, and inaccurate in cell separation. To overcome this problem, this research proposed an algorithm for cell separation based on enhanced edge detection. The objective of this study is identification of L1, L2, and L3, in touching cell based on edge detection and edge linking. Edge detection was performed on grayscale images and Hue images for the area of White Blood Cell segmentation results then the image results were combined. The classification of Acute Lymphoblastic Leukemia (ALL) cells was carried out using the geometry and texture features of cells and nucleus with Support Vector Machine (SVM) as the classifier. The dataset for the training process amounted to 668 ALL single cell images and for the testing process was 301 multi-cell ALL images. The testing results of the identification of ALL subtypes showed that the proposed cell separation method gave accuracy of 75.42%.
TL;DR: Machine learning methods were successfully carried out with optimized parameters for the classification of three levels of sweetness of pineapple, with the best accuracy is 82% using KNN with 3 neighbors.
Abstract: Electronic nose (e-nose) has been widely used to distinguish various scents in food. The output of e-nose is a signal that can be identified, compared, and analyzed. However, many researchers use e-nose without using standardization tools, therefore e-nose is still often questioned for its validity. This paper proposes an electronic nose (e-nose) to classify the sweetness of pineapples. The standard sweetness levels are measured by using a Brix meter as a standardization tool. The e-nose consists of a series of gas sensors MQ Series which are connected to the Arduino micro-controller. The sweetness levels measured by the Brix meter are then ordered into low, medium, high sweetness groups. These sweetness groups are used as label ground-truth for the e-nose. Signal processing PCA and mother wavelet is employed to reduce noise from the e-nose signals. The signal processing methods obtain optimal parameters to find the characteristics of each signal. Machine learning methods were successfully carried out with optimized parameters for the classification of three levels of sweetness of pineapple. The best accuracy is 82% using KNN with 3 neighbors.
TL;DR: The experimental results prove that the proposed FDCuT-DCT-SVD algorithm produces good imperceptibility and is also resistant to various types of attacks, including JPEG compression, noise enhancement attacks, filtering attacks, and other common attacks.
Abstract: One way to prevent image duplication is by applying watermarking techniques. In this work, the watermarking process is applied to medical images using the Fast Discrete Curvelet Transforms (FDCuT), Discrete Cosine Transform (DCT), and Singular Value Decomposition (SVD) methods. The medical image of the host is transformed using FDCuT so that three subbands are obtained. High Frequency (HF) subband selected for DCT and SVD applications. Meanwhile, SVD was also applied to the watermark image. The singular value on the host image is exchanged with the singular value on the watermark. Insertion of tears by exchanging singular values does not cause the quality of medical images to decrease significantly. The experimental results prove that the proposed FDCuT-DCT-SVD algorithm produces good imperceptibility. The proposed algorithm is also resistant to various types of attacks, including JPEG compression, noise enhancement attacks, filtering attacks, and other common attacks.
TL;DR: A new enhancement based on particle swarm optimization (PSO) algorithm called multiple inertia weight PSO (MIW- PSO) to solve the combined economic and emission load dispatch (CEELD) issues in the modern electrical power systems is proposed.
Abstract: Economic dispatch issues in power system aim to try getting an optimal plan for the power generators to minimize the fuel cost (FC) in parallel with satisfying system constraints. This paper proposes a new enhancement based on particle swarm optimization (PSO) algorithm called multiple inertia weight PSO (MIW-PSO) to solve the combined economic and emission load dispatch (CEELD) issues in the modern electrical power systems. Two electrical test systems are investigated in this study to validate the competence of the proposed algorithm. The obtained results for CEELD case using MIW-PSO compared with MOCPSO indicate a promising performance in terms of minimizing FC and pollutant emission (PE) are reduced 84.96 $/h and 12.01 kg/h for the first test system. As well as, for the second test system, compared with NSGA-RL are reduced 0.241 $/h and 3.15 kg/h. Moreover, the proposed algorithm has more accuracy, better convergence time, and higher quality solutions for the minimum CEELD compared with other methods.
TL;DR: This model is an unsupervised fall detector based on utilizing the deep learning technique to detect falls of the elderly people and shows that the SRAE model detects falls with high receiver operating characteristic area under curve (ROC AUC) and precision recall area under curves (PR AUC).
Abstract: This paper focuses on falling of the elderly people which is considered as one of the most critical issue that can face them in their life. To deal with such issue, we propose a new approach named a Spatio-temporal Residual AutoEncoder (SRAE) model. This model is an unsupervised fall detector based on utilizing the deep learning technique to detect falls of the elderly people. Our proposed model uses autoencoder based on convolutional neural network, convolutional long short term memory (ConvLSTM) network, and residual connections to extract spatial and temporal features of videos captured from thermal cameras. The reconstruction error of an autoencoder is used to detect falls recorded in such thermal videos. Furthermore, SRAE model is tested on the publicly available thermal dataset where thermal images conserve the privacy of the elderly under observation which is a very important issue. The obtained results show that the our proposed model detects falls with high receiver operating characteristic area under curve (ROC AUC) (97%) ,and precision recall area under curve (PR AUC) (93%) compared to denoising autoencoder (DAE), convolutional autoencoder (CAE), and convolutional long short term memory autoencoder (CLSTMAE) introduced in the literature.
TL;DR: Fuzzy Logic Based-on Sigmoid Membership Function was developed to improve lightness and contrast in coloured images and it had good average values for entropy, mean square error in saturation and lightness order error.
Abstract: Colour image enhancement plays an important role in image processing, computer vision and pattern recognition. Fuzzy logic techniques one of the methods used for digital image enhancement. In this study, Fuzzy Logic Based-on Sigmoid Membership Function (FLBSMF) was developed to improve lightness and contrast in coloured images. The FLBSMF algorithm was applied to the lightness component by using only the YIQ colour space, and the colour compounds were unchanged. The suggested algorithm was compared with other algorithms, such as fuzzy logic enhancement using membership function modification based on square operation, fuzzy logic based on histogram, histogram equalisation and multiscale retinex with colour restoration by using different criteria. Form the result the FLBSMF succeeded in enhancing colour images and it had good average values for entropy (7.44), mean square error in saturation (2.2E-08) and hue (8.11E-07), nature image quality (2.97) and lightness order error (0.90).
TL;DR: Artificial Neural Network (ANN) classifier outperformed all the classifiers under analysis with an accuracy of 93.00 % in predicting lung cancer and Artificial Neural Network classifier topped the list of classifiers in predicting infectious diseases such as hepatitis and dengue serotypes.
Abstract: Infectious and chronic diseases devastate millions of people across the world each year. Nonetheless, each type of disease substantiates differently. According to the National Centre for Health Statistics, USA, Infectious diseases or communicable diseases are the ones based on the cause, which spreads from person to person or animal to person caused by microorganisms such as bacteria or parasite and can be cured. Chronic diseases are based on the effect, which may have the origin of infectious disease, prolonged to three or more months, doesn’t spread from one person to another and cannot be cured. Some chronic diseases such as cervical cancer and liver cancer have originated from infectious diseases such as human papillomavirus (HPV) and hepatitis B, C virus. This paper focuses on various machine learning classification techniques in predicting chronic diseases such as Cardio Vascular Disease (CVD), Chronic Kidney Disease (CKD), lung cancer, and infectious diseases such as hepatitis and dengue serotypes. In the analysis, ABC4.5 classifier outperformed with accuracy of 92.76 % than the other classifiers in predicting Chronic Kidney Disease (CKD), Random Forest classifier achieved an accuracy of 90.32% which is higher than Logistic regression of accuracy 83.87% in predicting hepatitis. Hoeffding classifier achieves an accuracy of 88.56% which is higher than the other classifier in predicting Cardio Vascular Disease. Multi swarm optimized Multilayer perceptron achieved an accuracy of 85.18% which is higher than the particle swarmed optimized multilayer perceptron in predicting dengue serotypes. Artificial Neural Network (ANN) classifier outperformed all the classifiers under analysis with an accuracy of 93.00 % in predicting lung cancer.
TL;DR: The findings of the study show that the Reverse Engineering approach is the most efficient technique for analyzing complex malware.
Abstract: Nowadays, most of the cyber-attacks are initiated by extremely malicious programs known as Malware. Malwares are very vigorous and can penetrate the security of information and communication systems. While there are different techniques available for malware analysis, it becomes challenging to select the most effective approach. In this context, the decision-making process may be an efficient means of empirically assessing the impact of different methods for securing the web applications. In this research study, we have used a methodology that includes the integration of Fuzzy AHP and Fuzzy TOPSIS technique for evaluating the impact of different malware analysis techniques in web application perspective. This study uses different versions of a university’s web application for evaluating the impact of several existing malware analysis techniques. The findings of the study show that the Reverse Engineering approach is the most efficient technique for analyzing complex malware. The outcome of this study would definitely aid the future researchers and developers in selecting the appropriate techniques for scanning the web application code and enhancing the security.
TL;DR: A genetic algorithm and ant colony optimization modified with the greedy algorithm introduced to the system contains multiple heterogeneous embedded machines (HEMs) working as a cluster and outperforms the others by 33% more result quality.
Abstract: This paper presents a heuristic approach for workflow scheduling in heterogeneous distributed embedded system (HDES). A genetic algorithm (GA) and ant colony optimization (ACO) modified with the greedy algorithm introduced to the system contains multiple heterogeneous embedded machines (HEMs) working as a cluster. Users can remotely access and utilize their computational power. The communications on different types of buses are taken into account to find an optimal solution. New meta-heuristic information based on forwarding dependency is proposed to build probability for ACO to generate task priorities. Besides, a greedy algorithm for machine allocation is incorporated to complete task scheduling. Experiments based on random task graphs running in the HEM cluster demonstrate the effectiveness of the modified greedy ant colony optimization algorithm which outperforms the others by 33% more result quality.
TL;DR: To solve the problems of nonlinearity and power fluctuation linked on the Photovoltaic panel (PV) connected storage system and grid, because of the temperature and irradiation variation, the Maximum Power Point Tracking (MPPT) is obtained via the control of the duty cycle of DC/DC boost converter.
Abstract: To solve the problems of nonlinearity and power fluctuation linked on the Photovoltaic panel (PV) connected storage system and grid, because of the temperature and irradiation variation. We integrated three parts of the control. The first part, dedicated to developing an algorithm to eliminate the nonlinearity; therefore, we have obtained the Maximum Power Point Tracking (MPPT) via the control of the duty cycle of DC/DC boost converter. Consequently, to achieve the MPPT, we combined between tow algorithms Fuzzy Logic and Integral Backstepping (Fuzzy Logic-Integral Backstepping Controller) as a new strategy of MPPT. This MPPT based on the performances of the Fuzzy Logic Controller (FLC) as an estimator of the reference voltage, next, we apply the Integral Backstepping approach to generate the law control founded on the Lyapunov theory to augment the robustness and stability of the PV connected storage system and grid. Then, in the second part of the control, we add a Batteries Energy Storage System with a control management algorithm in the DC/DC side to eliminate any fluctuation of the output power of the PV system. In the third part of the control linked to the grid side, we use the three-phase Voltage Source Inverter Control (VSIC) as a charge controller for the stability of the grid parameters.
TL;DR: The central thesis of this paper is the implementation of an optimized Sliding Mode Control to a high power multiphase permanent magnet synchronous generator based direct-driven Wind Energy Conversion System (WECS), in order to attain an optimal regime characteristic.
Abstract: The central thesis of this paper is the implementation of an optimized Sliding Mode Control to a high power multiphase permanent magnet synchronous generator based direct-driven Wind Energy Conversion System (WECS), in order to attain an optimal regime characteristic. The principle of proposed approach is to control the Rotational Speed Dynamic to follow the desired value during wind variations. Moreover, the controller is designed to ensure an effective regulation of 4 dynamic models of Direct & Quadrature axis current instead of the regular hysteresis control due to its disadvantages. In this context, the main challenge of this research with a view to ensure the system stability, is to design a controller in a way to guarantee the consistency between the input of inner loop (4 axis current regulation) and the output of outer loop (speed regulation), taking into account the risk of declined effectiveness of closed-loop due to disappearing relative changes in control periods between the cascaded loops. Furthermore, the grid supply is managed using the proposed sliding mode approach in order to guarantee a power injection with Unity Power Factor. On the other hand, a challenging matter of pure SMC can be summed up in the produced chattering phenomenon. In this work, this issue has been mitigated by implementing a new smooth continuous switching control. In order to examine the suggested approach robustness and responsiveness over other techniques, a detailed analysis has been carried out under hard and random wind speed variations. The stability of improved SMC is verified using Lyapunov stability theory by taking into account system uncertainties to guarantee more efficacy. The suggested SMC is compared to a classical PI controller. The study results exhibit the excellent performance with high robustness of the improved SMC, by improving the system efficiency to 91,14%, compared to the PI with the ratio of 86,2%.
TL;DR: The proposed system is a combination of dense wavelength division multiplexing (DWDM) and multiple input multiple output (MIMO) techniques with Fork components and shows promising results in the performance and the quality of the received signal.
Abstract: Frees Space Optical (FSO) communication is one of the communication systems types that uses a free space to transmit information carried by light. It is previously known that the single input single output (SISO) technique is susceptible to atmospheric attenuation because of the effect of weather conditions. The proposed system is a combination of dense wavelength division multiplexing (DWDM) and multiple input multiple output (MIMO) techniques with Fork components. The main Fork component is configured in the Optisystem simulator to connect to another Fork components to increase the power of transmitted signals and ensure the arrival of the transmitted signals to the receiver of the FSO system. The key idea in the proposed system is to mitigate the attenuation that will be happened to the transmitted optical signal which propagated through free space due to different weather conditions. The study is done based on using the Optisystem simulation toolbox that is used to emulate different weather attenuation conditions in two types of FSO systems. The first is the traditional DWDMSISO and the second is a proposed system named Hybrid DWDM-MIMO. A comparison between the proposed system and the traditional system is made in terms of the quality factor under different weather attenuation. The proposed system shows promising results in the performance and the quality of the received signal. The transmission path length of the proposed system under dense fog attenuation of 260 dB/km is enhanced to (30.43%) in comparison with traditional DWDM-SISO, also transmission path length of the proposed system under heavy rain attenuation of 9.29 dB/km is enhanced to (55.55%) in comparison with traditional DWDMSISO. The transmission path length of the proposed system under heavy dry snow of attenuation of 131.835 dB/km is enhanced to (26.19%) in comparison with traditional DWDM-SISO.
TL;DR: A Neural Style Geometric Transformation (NSGT) is proposed as a data augmentation technique for Balinese carvings recognition by combining Neural Style Transfers and Geometric Transformations for a small dataset solution.
Abstract: The preservation of Balinese carving data is a challenge in recognition of Balinese carving. Balinese carvings are a cultural heritage found in traditional buildings in Bali. The collection of Balinese carving images from public images can be a solution for preserving cultural heritage. However, the lousy quality of taking photographs, e.g., skewed shots, can affect the recognition results. Research on the Balinese carving recognition has existed but only recognizes a predetermined image. We proposed a Neural Style Geometric Transformation (NSGT) as a data augmentation technique for Balinese carvings recognition. NSGT is combining Neural Style Transfers and Geometric Transformations for a small dataset solution. This method provides variations in color, lighting, rotation, rescale, zoom, and the size of the training dataset, to improve recognition performance. We use MobileNet as a feature extractor because it has a small number of parameters, which makes it suitable to be applied on mobile devices. Eight scenarios were tested based on image styles and geometric transformations to get the best results. Based on the results, the proposed method can improve accuracy by up to 16.2%.
TL;DR: Results show that data with improved quality of the images will improve accuracy for fish species identification and improvement using the proposed method of 3.56%.
Abstract: Automatic identification of fish species is very complex and challenging because of the low quality of the marine environment. Thus, the identification of fish species using computer vision technology is disrupted. However, various researchers only focus on determining the best fish identification method without considering the quality of the data used. Therefore, this study presented a new workflow in identifying fish species. A combination of feature extraction methods and a backpropagation neural network (BPNN) method was used, which was based on image quality improvement techniques using contrast limited adaptive histogram equalization (CLAHE) with adaptive threshold by fuzzy c-means. This study compared the results of fish identification on the original data and image data that were enhanced using several classifications of machine learning. The results show that data with improved quality of the images will improve accuracy for fish species identification and improvement using the proposed method of 3.56%. This could support the reduction of invasive fish populations through automated fish identification systems in unrestricted natural environments based on computer vision technology.
TL;DR: The proposed filter-based improved decision tree sentiment classification model for real-time amazon product review data recommends the product based on the user query by prediction using a new novel normalized product review sentiment score and ranked feature selection measure.
Abstract: E-Commerce product features and reviews are considered to be the essential factors in real-time e-commerce sites for product recommendation systems. Due to inaccuracy decision patterns, in most cases e-commerce user fails to predict the products based on the user ratings and review comments. Traditional sentiment classification models are independent of data filtering, transformation and sentiment score computing techniques which require high computing memory, time and mostly leading to false-positive rate. To overcome these issues, a novel sentiment score-based product recommendation model is proposed on the real-time product data. In this model, a new product ranking score, filtering, and hybrid decision tree classifiers are implemented. Initially, real-time amazon product review data is captured using Document Object Model (DOM) parser. The features from the review comments are extracted using lexicon Feature Dictionary (FD) and AFINN, Normalized Product Review Score (NPRS) are generated to compute the class label for product review sentiment prediction. Ranked Principal Component Analysis (RPCA) is used as a feature selection measure to overcome the problem of data sparsity. Random Tree, Hoeffding Tree, Adaboost + Random Tree, the three variants of decision tree classifiers are used for product sentiment classification. The proposed filter-based improved decision tree sentiment classification model for real-time amazon product review data recommends the product based on the user query by prediction using a new novel normalized product review sentiment score and ranked feature selection measure. The proposed product recommendation, the decision-making system maximizes sentiment classification accuracy. Experimental results are compared against the traditional decisionmaking classification models in terms of correctly classified instances, error rate, and PRC, F-measure, kappa statistics. The proposed model experimental results show high efficiency.