Search

In a smart grid, the main goals are to provide grid stability, improve power system performance and security, and reduce operations, system maintenance, and planning costs. The prediction stability of smart grid (SG) systems is essential in terms of power loss minimization and the importance of adequate energy policies. SG systems must accurately predict the energy demand and ensure the right amount of energy is available at the right time. If the prediction is inaccurate, it can lead to costly energy production or usage errors and create considerable inefficiencies in the power grid. Due to this, this manuscript offers five different cascade methods to detect the stability of SG systems. Detecting the stability of SG systems enables the grid to respond quickly to changes in demand and supply, improves system reliability, reduces power outages, and increases the overall efficiency of the grid. The present work proposed five different cascade methods with pre-processing, training and testing division, and the classification stages of the classification procedure for estimating SG stability. In the first pre-processing stage, the SG dataset is pre-proceeded with the feature selection (Relief, Fast Correlation-Based Filter (FCBF), and supervised attribute filter). The resampling (the bootstrapping), the Fuzzy C-Means Clustering-Based Feature Weighting (FCMFW), the resampling then feature selection (supervised attribute filter), and the feature selection (supervised attribute filter), then FCMFW. In the second stage, the training and testing division stage, the SG dataset was separated into three test and training data methods before the classification algorithm: The 5 Fold Cross Validation (FVC), 10 FVC, and hold-out (50–50%). In the third stage, the classification stage, five different classification algorithms, including Naive Kernel Bayes, Linear Support Vector Machine (SVM), Weighted K-Nearest Neighbors, Begged Trees, and Narrow Neural Network classifying algorithms, are used to classify the SG dataset. The simulation results of this study demonstrated that the suggested cascade ML system had achieved significant accuracy in predicting SG stability. The best cascade method is the feature selection (supervised attribute filter) + FCMFW + 10 FCV and then performing the bagged trees algorithm; thus, the new approach affords an accuracy of 99.9%. Furthermore, due to the rapid growth of ML techniques, sensors, and smart meters technologies, with Machine to Machine communication via the internet of things (IoT), the real-time identification process is made practical with higher accuracy. For this reason, our future research will focus on an IoT-based SG system, an E-stability determination system. Thanks to the proposed cascade method, the SG dataset can be classified easily, quickly, and reliably. E-stability determination systems can help to fast detect, predict, and respond, which is an important application of IoT on the grid systems. [ABSTRACT FROM AUTHOR]

Epilepsy is one of the most common neurological diseases, affecting approximately 50 million people. This illness can be diagnosed by electroencephalogram (EEG), whose analysis depends on human interpretation, which may lead to divergent results and be imprecise and error-prone. Moreover, one estimates more than 80% of the epilepsy examinations return no anomalies at all, wasting the effort of analysis. In this context, machine learning (ML) methods are used to aid in epilepsy detection. An essential task for classifier building is feature extraction, for which there are many measures that can be computed. However, part of these features can be irrelevant or hinder the performance of ML methods. This paper proposes an automatic way to classify EEG segments by using a reduced set of features. The proposed method combines multispectral analysis, feature selection, and classifier building approaches. As our main contribution, a methodology to minimize the number of features for classifier building is proposed. A total of 285 measures are extracted. Afterward, two attribute selection approaches are used: Pearson's correlation coefficient filter and wrapper based on the genetic algorithm. Then, five well-known ML methods (k-nearest-neighbor, support vector machine, naive Bayes, artificial neural network, and random forest) were used to build 281 different classifiers. As a result, the proposed classifiers reached an accuracy between 87.2 and 90.99% and considerably reduced the number of features from 285 to 30, keeping competitive scores. Additionally, statistical hypothesis tests prove that our proposed approach is as efficient as using the complete feature dataset for classifier building. [ABSTRACT FROM AUTHOR]

In this manuscript, imbalanced and small sample space (IB-SSS) dataset problems for pedestrian gender classification using fusion of selected deep and traditional features (PGC-FSDTF) are considered. In this regard, data preparation is first done through data augmentation and preprocessing steps to handle imbalanced classification problem and environmental effects, respectively. The proposed approach follows different types of feature extraction schemes, for instance, pyramid histogram of oriented gradients, hue saturation value histogram, deep visual features of DenseNet201 and InceptionResNetV2-based convolutional neural network architectures. The parallel fusion method computes the maximum and average values-based features from the learned features of both deep networks. Features are selected through features selection methods such as entropy and principal component analysis (PCA). The subsets of features are serially fused and provided to multiple classifiers to perform gender classification on IB-SSS datasets. Resultantly, the proposed PGC-FSDTF method shows better results in terms of different accuracies (overall, mean, and balanced), and area under curve on selected datasets. Further, improved results are achieved on applied datasets using PCA-based selected features and medium Gaussian support vector machine (M-SVM) classifier. These results on different datasets confirm that the selected feature combination provides a way to handle IB-SSS issues for PGC effectively. [ABSTRACT FROM AUTHOR]

Discovering a hearing disorder at an earlier intervention is critical for reducing the effects of hearing loss and the approaches to increase the remaining hearing ability can be implemented to achieve the successful development of human communication. Recently, the explosive dataset features have increased the complexity for audiologists to decide the proper treatment for the patient. In most cases, data with irrelevant features and improper classifier parameters causes a crucial influence on the audiometry system in terms of accuracy. This is due to the dependent processes of these two, where the classification accuracy performance could be worsened if both processes are conducted independently. Although the filter algorithm is capable of eliminating irrelevant features, it still lacks the ability to consider feature reliance and results in a poor selection of significant features. Improper kernel parameter settings may also contribute to poor accuracy performance. In this paper, an ensemble filters feature selection based on Information Gain (IG), Gain Ratio (GR), Chi-squared (CS), and Relief-F (RF) with harmonize optimization of Particle Swarm Optimization (PSO) and Support Vector Machine (SVM) is presented to mitigate these problems. Ensemble filters are utilized so that the initial top dominant features relevant for classification can be considered. Then, PSO and SVM are optimized simultaneously to achieve the optimal solution. The results on a standard Audiology dataset show that the proposed method produces 96.50% accuracy with optimal solution compared to classical SVM, which signifies the proposed method is effective in handling high dimensional data for hearing disorder prediction. [ABSTRACT FROM AUTHOR]

Non-motor manifestations of Parkinson's disease (PD) appear early and have a significant impact on the quality of life of patients, but few studies have evaluated their predictive potential with machine learning algorithms. We evaluated 9 algorithms for discriminating PD patients from controls using a wide collection of non-motor clinical PD features from two databases: Biocruces (96 subjects) and PPMI (687 subjects). In addition, we evaluated whether the combination of both databases could improve the individual results. For each database 2 versions with different granularity were created and a feature selection process was performed. We observed that most of the algorithms were able to detect PD patients with high accuracy (>80%). Support Vector Machine and Multi-Layer Perceptron obtained the best performance, with an accuracy of 86.3% and 84.7%, respectively. Likewise, feature selection led to a significant reduction in the number of variables and to better performance. Besides, the enrichment of Biocruces database with data from PPMI moderately benefited the performance of the classification algorithms, especially the recall and to a lesser extent the accuracy, while the precision worsened slightly. The use of interpretable rules obtained by the RIPPER algorithm showed that simply using two variables (autonomic manifestations and olfactory dysfunction), it was possible to achieve an accuracy of 84.4%. Our study demonstrates that the analysis of non-motor parameters of PD through machine learning techniques can detect PD patients with high accuracy and recall, and allows us to select the most discriminative non-motor variables to create potential tools for PD screening. [ABSTRACT FROM AUTHOR]

Biological data generally contain complex and high-dimensional samples. In addition, the number of samples in biological datasets is much fewer than the number of features, so the vast number of features should be selected carefully and determine the optimal subset of features. Feature selection (FS) is a vital stage in biological data mining applications (e.g., classification) for dealing with the curse of dimensionality problems and finding highly informative features. This work proposes an effective FS approach based on a new version of Gray Wolf Optimizer (GWO) called Multi-strategy Gray Wolf Optimizer (MSGWO) for better features selection for biological data classification. The use of MSGWO in feature selection is to find the optimal subset of features between classes, solve premature convergence, and enhance the local search ability of the GWO algorithm. Multiple exploration and exploitation strategies are proposed to enhance the global search and local search abilities of the GWO algorithm through the optimization process. The support vector machine (SVM) classifier is used to evaluate the proposed GWO-based FS approaches. MSGWO was evaluated on thirteen high-dimensional biological datasets obtained from the UCI repository with a smaller number of instances. The reported results confirm that employing multiple exploration and multiple exploitation strategies is highly useful for enhancing the search tendency of the MSGWO in the FS domain. Statistical tests proved that the superiority of the proposed approach is statistically significant as compared to the basic GWO and similar wrapper-based FS techniques, including binary particle swarm optimization (BPSO), binary bat algorithm (BBA), binary gravitational search algorithm (BGSA), and binary whale optimization algorithm (BWOA). In terms of classification accuracy, MSGWO yielded better accuracy rates than the standard GWO algorithm on 84% of applied biological datasets. MSGWO also recorded better accuracy rates than its other competitors in all 13 cases. In terms of the lowest number of selected features, MSGWO yielded excellent reduction rates compared to its peers. [ABSTRACT FROM AUTHOR]

This study focuses on using EEG signal-based behavioral biometric data to classify and identify persons. A person identification system based on a nonlinear model with excellent estimation accuracy was proposed. Six tasks were performed on all subjects utilizing EEG data collected from 64 channels and 96 subjects. The stages of this model included feature creation, feature selection, and classification. To generate features from EEG signals, statistical methods were used. Using neighborhood component analysis, the best 15 features were chosen. The selected feature vector was given as input to deep feature neural network (DNN) model and traditional ML classifier decision tree (DT), k-nearest neighbor (KNN), support vector machine (SVM), ensemble—random forest (RF) classifiers. The classification performance of the proposed methods for six tasks was separately tested and compared. The average accuracy of DT, RF, KNN, SVM and DNN classifiers for all tasks was 98.63%, 100%, 99.96%, 99.91%, and 100%, respectively. The results show that RF and DNN classifier-based models are highly effective in the classification and person identification using EEG signals. [ABSTRACT FROM AUTHOR]

The experts' decisions and evaluating the patients' data are the most significant parts affecting the breast cancer analysis. For early breast cancer detection, numerous techniques of machine learning not only can assist in examining and diagnosis the medical data quickly but also decrease the potential errors that could be occurred due to inexpert or unskilled decision-makers. Support vector machine is one of the famous classifiers that has already made an important contribution to the field of cancer classification. However, configurations of different kernel function and their parameters can significantly affect the performance of the SVM classifier. To further improve the classification accuracy of the SVM classifier for breast cancer diagnosis, an intelligent cancer classification method is proposed based on selecting a feature subset and optimizing the relevant parameters (i.e., penalty factor parameter (c ) and kernel parameter γ ) of the SVM classifier concurrently through an intelligent algorithm using the Jaya algorithm. Then, this method (Jaya-SVM) was applied to precisely characterize the breast cancer dataset, including 699 samples, which are 458 and 241 for benign and malignant, respectively. Furthermore, to evaluate the effectiveness of the proposed Jaya-SVM classifier, it is compared in terms of the computational complexity and the classification accuracy with several combinatorial metaheuristic classifiers, namely the genetic algorithm (GA), differential evolution (DE), particle swarm optimization (PSO), and cuckoo search (CS) based-SVM. Apart from this, a Breast Cancer Coimbra Dataset taken from the UCI library is used to validate the effectiveness of the proposed method. The results are presented, explained, and conclusions are drawn. [ABSTRACT FROM AUTHOR]

Cancer is a severe condition of uncontrolled cell division that results in a tumor formation that spreads to other tissues of the body. Therefore, the development of new medication and treatment methods for this is in demand. Classification of microarray data plays a vital role in handling such situations. The relevant gene selection is an important step for the classification of microarray data. This work presents gene encoder, an unsupervised two-stage feature selection technique for the cancer samples' classification. The first stage aggregates three filter methods, namely principal component analysis, correlation, and spectral-based feature selection techniques. Next, the genetic algorithm is used, which evaluates the chromosome utilizing the autoencoder-based clustering. The resultant feature subset is used for the classification task. Three classifiers, namely support vector machine, k-nearest neighbors, and random forest, are used in this work to avoid the dependency on any one classifier. Six benchmark gene expression datasets are used for the performance evaluation, and a comparison is made with four state-of-the-art related algorithms. Three sets of experiments are carried out to evaluate the proposed method. These experiments are for the evaluation of the selected features based on sample-based clustering, adjusting optimal parameters, and for selecting better performing classifier. The comparison is based on accuracy, recall, false positive rate, precision, F-measure, and entropy. The obtained results suggest better performance of the current proposal. [ABSTRACT FROM AUTHOR]

In medical science, collecting and classifying data from various diseases is a vital task. The confused and large amounts of data are problems that prevent us from achieving acceptable results. One of the major problems for diabetic patients is a failure to properly diagnose the disease. As a result of this mistake in diagnosis or failure in early diagnosis, the patient may suffer from complications such as blindness, kidney failure, and cutting off the toes. Nowadays, doctors diagnose the disease by relying on their experience and knowledge and performing complex and time-consuming tests. One of the problems with current diabetic, diagnostic methods is the lack of appropriate features to diagnose the disease and consequently the weakness in its diagnosis, especially in its early stages. Since diabetes diagnosis relies on large amounts of data with many parameters, it is necessary to use machine learning methods such as support vector machine (SVM) to predict the complications of diabetes. One of the disadvantages of SVM is its parameter adjustment, which can be accomplished using metaheuristic algorithms such as particle swarm optimization algorithm (PSO), genetic algorithm, or grey wolf optimizer (GWO). In this paper, after preprocessing and preparing the dataset for data mining, we use SVM to predict complications of diabetes based on selected parameters of a patient acquired by laboratory test using improved GWO. We improve the selection process of GWO by employing dynamic adaptive middle filter, a nonlinear filter that assigns appropriate weight to each value based on the data value. Comparison of the final results of the proposed algorithm with classification methods such as a multilayer perceptron neural network, decision tree, simple Bayes, and temporal fuzzy min–max neural network (TFMM-PSO) shows the superiority of the proposed method over the comparable ones. [ABSTRACT FROM AUTHOR]

A Gaussian based Particle Swarm Optimization Gravitational Search Algorithm (GPSOGSA) is being proposed for extensive feature selection that serves highly in making effective predictions. GPSOGSA helps to overcome the problem of being stuck into the local optima and influences the local searching ability, thus it aims to bridge the gap of exploration and exploitation. The algorithm also limits the usage of too many parameters like acceleration factors, maximum velocity, inertia weight that plays a vital role in PSO, GSA and PSOGSA. The efficacy of the algorithm has been tested upon unimodal and multimodal benchmark functions. We have also evaluated the performance of the algorithm by applying it on various benchmark datasets. The algorithm uses a wrapper-based approach that includes Support Vector Machine as a learner algorithm, and improves both the execution time and the performance accuracy. The findings show that the proposed algorithm could escape from local optimum and converges faster than the PSO, GSA and PSOGSA algorithms. [ABSTRACT FROM AUTHOR]

The quest to improve performance accuracy and prediction speed in machine learning algorithms cannot be overemphasized, as the need for machines to outperform humans continue to grow. Accordingly, several studies have proposed methods to improve prediction performance and speed particularly for spatio-temporal analysis. This study proposes a novel classifier that leverages ensemble techniques to improve prediction performance and speed. The proposed classifier, Ada-AdaSVM uses an AdaBoost feature selection algorithm to select small features of input datasets for a joint support vector machine (SVM)–AdaBoost classifier. The proposition is evaluated against a selection of existing classifiers (SVM, AdaSVM and AdaBoost) using the Jaffe, Yale, Taiwanese facial expression database (TFEID) and CK + 48 datasets with Haar features as the preferred method for feature extraction. The findings indicated that Ada-AdaSVM outperforms SVM, AdaSVM and AdaBoost classifiers in terms of speed and accuracy. [ABSTRACT FROM AUTHOR]

With the exponential growth of the amount of data being generated, stored and processed on a daily basis in the machine learning, data analytics and decision-making systems, the data preprocessing established itself as the key factor for building reliable high-performance machine learning models. One of the roles in preprocessing is variable reduction using feature selection methods; however, the processing time needed for these methods is a major drawback. This study aims at mitigating this problem by migrating the algorithm to a MapReduce implementation suitable for parallelization on a high number of commodity hardware units. The genetic algorithm-based methods were put in the focus of this study. Hadoop, an open-source MapReduce library, was used as a framework for implementing parallel genetic algorithms within our research. The representative machine learning methods, SVM (support vector machine), ANN (artificial neural network), RT (random tree), logistic regression and Naive Bayes, were embedded into implementation for feature selection. The feature selection methods were applied to four NSL-KDD data sets, and the number of features is reduced from cca 40 to cca 10 data sets with the accuracy of 90.45%. These results have both significant practical and theoretical impact. On the one hand, the genetic algorithm has been parallelized in the MapReduce manner, which has been considered unachievable in a strict sense. Furthermore, the genetic algorithm allows randomness-enhanced feature selection and its parallelization reduces overall data preprocessing and allows larger population count which in turn leads to better feature selection. On the practical side, it has been shown that this implementation outperforms the existing feature selection methods. [ABSTRACT FROM AUTHOR]

Software testing process is a crucial part in software development. Generally the errors made by developers get fixed at a later stage of the software development process. This increases the impact of the defect. To prevent this, defects need to be predicted during the initial days of the software development, which in turn helps in efficient utilization of the testing resources. Defect prediction process involves classification of software modules into defect prone and non-defect prone. This paper aims to reduce the impact of two major issues faced during defect prediction, i.e., data imbalance and high dimensionality of the defect datasets. In this research work, various software metrics are evaluated using feature selection techniques such as Recursive Feature Elimination (RFE), Correlation-based feature selection, Lasso, Ridge, ElasticNet and Boruta. Logistic Regression, Decision Trees, K-nearest neighbor, Support Vector Machines and Ensemble Learning are some of the algorithms in machine learning that have been used in combination with the feature extraction and feature selection techniques for classifying the modules in software as defect prone and non-defect prone. The proposed model uses combination of Partial Least Square (PLS) Regression and RFE for dimension reduction which is further combined with Synthetic Minority Oversampling Technique due to the imbalanced nature of the used datasets. It has been observed that XGBoost and Stacking Ensemble technique gave best results for all the datasets with defect prediction accuracy more than 0.9 as compared to algorithms used in the research work. [ABSTRACT FROM AUTHOR]

Features subset selection was commonly used in data mining and artificial intelligence techniques to produce a model with a minimal set of features that enhances the performance of the classifier. The essential motive for selecting features is to avoid the problem of a number of dimensions trap. This paper introduces a new technique of selection of features dependent on the modified of binary teaching–learning-based optimization and the suggested method called MBTLBO. This algorithm (teaching learning-based optimization TLBO) is one of the present metaheuristic that is been widely utilized to a several of intractable optimization issues in recent times. Such algorithm has been combined with supervised data mining technique (support vector machine) for the implementation of feature subset selection problem in binary identification. The collection of specific risk features with the rheumatic disease was implemented. The findings revealed that the new approach (MBTLBO) increases the accuracy of classification. [ABSTRACT FROM AUTHOR]

A software defect report is a bug in the software system that developers and users submit to the software defect library during software development and maintenance. Managing a software defect report that is overwhelming is a challenging task. The traditional method is manual identification, which is time-consuming and laborious and delays the repair of important software defects. Based on the above background, the purpose of this paper is to study the software defect prediction (SDP) model based on LASSO–SVM. In this paper, the problem of poor prediction accuracy of most SDP models is proposed. A SDP model combining minimum absolute value compression and selection method and support vector machine algorithm is proposed. Firstly, the feature selection ability of the minimum absolute value compression and selection method is used to reduce the dimension of the original data set, and the data set not related to SDP is removed. Then, the optimal value of SVM is obtained by using the parameter optimization ability of cross-validation algorithm. Finally, the SDP is completed by the nonlinear computing ability of SVM. The accuracy of simulation results is 93.25% and 66.67%, recall rate is 78.04%, and f-metric is 72.72%. The results show that the proposed defect prediction model has higher prediction accuracy than the traditional defect prediction model, and the prediction speed is faster. [ABSTRACT FROM AUTHOR]

In order to provide benchmark performance for Urdu text document classification, the contribution of this paper is manifold. First, it provides a publicly available benchmark dataset manually tagged against 6 classes. Second, it investigates the performance impact of traditional machine learning-based Urdu text document classification methodologies by embedding 10 filter-based feature selection algorithms which have been widely used for other languages. Third, for the very first time, it assesses the performance of various deep learning-based methodologies for Urdu text document classification. In this regard, for experimentation, we adapt 10 deep learning classification methodologies which have produced best performance figures for English text classification. Fourth, it also investigates the performance impact of transfer learning by utilizing Bidirectional Encoder Representations from Transformers approach for Urdu language. Fifth, it evaluates the integrity of a hybrid approach which combines traditional machine learning-based feature engineering and deep learning-based automated feature engineering. Experimental results show that feature selection approach named as normalized difference measure along with support vector machine outshines state-of-the-art performance on two closed source benchmark datasets CLE Urdu Digest 1000k, and CLE Urdu Digest 1Million with a significant margin of 32% and 13%, respectively. Across all three datasets, normalized difference measure outperforms other filter-based feature selection algorithms as it significantly uplifts the performance of all adopted machine learning, deep learning, and hybrid approaches. The source code and presented dataset are available at Github repository https://github.com/minixain/Urdu-Text-Classification. [ABSTRACT FROM AUTHOR]

This paper presents a content-based image retrieval technique that focuses on extraction and reduction in multiple features. To obtain multi-level decomposition of the image by extracting approximation and correct coefficients, discrete wavelet transformation is applied to the RGB channels initially. Therefore, both approximation and correct coefficients are applied to the dominant rotated local binary pattern termed as texture descriptor which is computationally effective and rotationally invariant. For a local neighbor patch, a rotation invariance function image is obtained by measuring the descriptor relative to the reference. The proposed approach contains the complete structural information extracted from the local binary patterns and also extracts the additional information using the information of magnitude, thereby achieving extra discriminative power. Then, GLCM description is used by obtaining the dominant rotated local binary pattern image to extract the statistical characteristics for texture image classification. The proposed technique is applied to CORAL dataset with the help of particle swarm optimization-based feature selector to minimize the number of features that can be used during the classification process. The three classifiers, i.e., support vector machine, K-nearest neighbor, and decision tree, are trained and tested. The comparison is based in terms of Accuracy, precision, recall, and F-measure performance metrics for classification. Experimental results show that the proposed approach achieves better accuracy, precision, recall, and F-measure values for most of the CORAL dataset classes. [ABSTRACT FROM AUTHOR]

Intelligent recognition sports management system applied to athletes' sports management system can effectively improve the quality of training and competition. At present, the technical bottlenecks of most sports management systems are in the sports recognition classification module. Based on BP neural network, this research has carried out experiments and analysis from feature extraction, feature selection, principal component analysis to the selection of support vector machine model functions. Moreover, this study designs a human pose recognition algorithm based on a convolutional neural network, optimizes the structure and parameters of the convolutional neural network, and designs a motion management system based on this algorithm. In addition, this study compares the performance of the proposed algorithm with other algorithms. The research shows that the method proposed in this research has certain practical significance and can provide theoretical references for subsequent related research. [ABSTRACT FROM AUTHOR]

Person re-identification (ReID) is an imperative area of pedestrian analysis and has practical applications in visual surveillance. In the person ReID, the robust feature representation is a key issue because of inconsistent visual appearances of a person. Also, an exhaustive gallery search is required to find the target image against each probe image. To answer such issues, this manuscript presents a framework named features-based clustering and deep features in person ReID. The proposed framework initially extracts three types of handcrafted features on the input images including shape, color, and texture for feature representation. To acquire optimal features, a feature fusion and selection technique is applied to these handcrafted features. Afterward, to optimize the gallery search, features-based clustering is performed for splitting the whole gallery into k consensus clusters. For relationship learning of gallery features and related labels of the chosen clusters, radial basis kernel is employed. Later on, cluster-wise, images are selected and provided to the deep convolution neural network model to obtain deep features. Then, a cluster-wise feature vector is obtained by fusing the deep and handcrafted features. It follows the feature matching process where multi-class support vector machine is applied to choose the related cluster. Finally, to find accurate matching pair from the classified cluster(s) instead of the whole gallery search, a cross-bin histogram-based distance similarity measure is used. The recognition rate at rank 1 is attained as 46.82%, 48.12%, and 40.67% on selected datasets VIPeR, CUHK01, and iLIDS-VID, respectively. It confirms the proposed framework outperforms the existing ReID approaches. [ABSTRACT FROM AUTHOR]

The intrusion detection is an essential section in network security because of its immense volume of threats which bothers the computing systems. The real-time intrusion detection dataset comprises redundant or irrelevant features. The duplicate features make it quite challenging to locate the patterns for intrusion detection. Hybrid optimization scheme (HOS) is designed for combining adaptive artificial bee colony (AABC) with adaptive particle swarm optimization (APSO) for detecting intrusive activities. The schemes are aggregated for locating improved optimization-based outcomes, and the precision during categorization is acquired using tenfold cross-validation scheme. The main objective of the proposed method is to improve the rate of precision in intrusion activities in internetwork by choosing the relevant features. Effectiveness of the hybrid categorization scheme is accessed using an NSL-KDD dataset. Single feature selection method and random feature selection method are used to assess the proposed HOS intrusion detection approaches. The effectiveness of the designed scheme is evaluated with existing machine learning schemes such as Naive Bayes, AABC, APSO, and support vector machine, which outperform the HOS. [ABSTRACT FROM AUTHOR]

In microarray experiments, the sample size is considerably smaller than that of the feature size, thereby imposing the curse of dimensionality problem. To resolve this issue, evolutionary algorithms are often utilized. In this paper, a novel framework for feature selection and classification of the microarray data is presented. Initially, a statistical filter, namely ANOVA, is used to select the relevant genes (features) from the original set of genes. Then, an evolutionary wrapper-based approach utilizing the principles of enhanced Jaya (EJaya) algorithm and forest optimization algorithm (FOA) is proposed to find the optimal set of genes from the previously selected genes. The main objective of using EJaya is to tune the two important parameters, namely local seeding changes and global seeding changes of FOA. During the selection of the optimal set of genes, support vector machine is employed as a classifier to classify the microarray data. To perform a comprehensive experimental study, the proposed method is tested on both binary-class and multi-class microarray datasets. From the extensive result analysis, it has been observed that the proposed technique achieves better classification accuracy with considerably less number of features than that of the benchmark schemes. [ABSTRACT FROM AUTHOR]

This paper proposes an improved computer-aided diagnosis model to identify mammographic images as normal or abnormal, and further, benign or malignant. The proposed scheme employs all the steps associated with other classification schemes; however, the contribution of the suggested scheme is fourfold. Initially, a fusion-based feature extraction method is employed to obtain the features using a combination of 2-D block discrete wavelet transform (2D-BDWT) and gray-level co-occurrence matrix (GLCM). Next, principal component analysis (PCA) is utilized to reduce the large dimension of the feature vector. Furthermore, to select the most optimal features from the reduced set of features, forest optimization algorithm (FOA) is used. The FOA-based feature selection algorithm is utilized as a wrapper-based technique which includes both feature selection and classification. In the proposed framework, several classifiers, namely support vector machine (SVM), k-nearest neighbor (k-NN), and decision tree (C4.5), are applied. The proposed method is compared with the benchmark schemes on two standard datasets, namely Mammographic Image Analysis Society (MIAS) and Digital Database for Screening Mammography (DDSM). Simulation results and analysis confirm that the proposed scheme brings potential improvements with respect to classification accuracy, sensitivity, specificity, area under curve, F-score, and Matthews correlation coefficient. The classification accuracy is measured with respect to normal versus abnormal and further, benign versus malignant. The proposed scheme with different combinations of classifiers, namely 2D-BDWT + GLCM + PCA + FOA + SVM, 2D-BDWT + GLCM + PCA + FOA + k-NN, and 2D-BDWT + GLCM + PCA + FOA + C4.5, achieves a maximum classification accuracy of 100% for both the MIAS and DDSM datasets. [ABSTRACT FROM AUTHOR]

Remote sensing image classification plays an important role in a wide range of applications and has caused widely concerns. During the last few years, great efforts have been made to develop a number of scene classification methods for remote sensing images. However, the existing remote sensing image classification methods do not perform satisfactorily in dealing with multi-class classification problems and rely heavily on the quality of data sets. These disadvantages seriously restrict the application of remote sensing image, including industrial research, analysis and calculation of land use and land coverage. To this end, this paper proposes a remote sensing image classification algorithm based on the sparse regularized feature learning method. Specifically, after constructing bag of features by using speeded up robust features extraction algorithm, direct sparsity optimization-based feature selection method is applied for selecting discriminative features, which is used for constructing support vector machine classifier model. The proposed algorithm has been evaluated and compared with other advanced feature selection methods on four public remote sensing image data sets. The experimental results demonstrate the effectiveness of our proposed image classification algorithm, which has been successfully applied to remote sensing image classification tasks. [ABSTRACT FROM AUTHOR]

With the development of science and technology, the classification of polarimetric SAR images has become an important part of the research of target recognition and image interpretation. However, for the research method is relatively simple and the accuracy is low, this paper carries out the work from the two aspects of feature extraction and feature classification of the ground object, and analyzes and studies the application and value of the polarimetric SAR system. The basic algorithm of polarization SAR image classification is proposed. A polarimetric SAR image feature classification method based on polarization target decomposition and support vector machine is proposed. Four kinds of scattering features and Freeman decomposition are obtained by Cloude decomposition. The simulation results show that the accuracy of using combined features is about 6.5% higher than that of single features. A polarization classification model based on polarization target decomposition and limit learning method is proposed. The simulation experiment shows ELM learning. The algorithm is indeed much faster than SVM learning. In this paper, a polarimetric SAR image classification method based on improved scattering mechanism coefficients is proposed, and the effectiveness of the polarimetric SAR image classification method based on improved scattering mechanism coefficients is verified. Experimental results show that after feature selection, the method of combining Freeman decomposition and Wishart classifier can get better classification results. [ABSTRACT FROM AUTHOR]

One critical issue within network security refers to intrusion detection. The nature of intrusion attempts appears to be nonlinear, wherein the network traffic performance is unpredictable, and the problematic space features are numerous. These make intrusion detection systems (IDSs) a challenge within the research arena. Hence, selecting the essential aspects for intrusion detection is crucial in information security and with that, this study identified the related features in building a computationally efficient and effective intrusion system. Accordingly, a modified feature selection (FS) algorithm called modified binary grey wolf optimisation (MBGWO) is proposed in this study. The proposed algorithm is based on binary grey wolf optimisation to boost the performance of IDS. The new FS algorithm selected an optimal number of features. In order to evaluate the proposed algorithm, the benchmark of NSL-KDD network intrusion, which was modified from 99-data set KDD cup to assess issues linked with IDS, had been applied in this study. Additionally, the support vector machine was employed to classify the data set effectively. The proposed FS and classification algorithms enhanced the performance of the IDS in detecting attacks. The simulation outcomes portrayed that the proposed algorithm enhanced the accuracy of intrusion detection up to 99.22% and reduction in the number of features from 41 to 14. [ABSTRACT FROM AUTHOR]

Spam detection on social networks is increasingly important owing to the rapid growth of social network user base. Sophisticated spam filters must be developed to deal with this complex problem. Traditional machine learning approaches such as neural networks, support vector machines and Naïve Bayes classifiers are not effective enough to process and utilize complex features present in high-dimensional data on social network spam. Moreover, the traditional objective criteria of social network spam filters cannot cope with different costs assigned to type I and type II errors. To overcome these problems, here we propose a novel cost-sensitive approach to social network spam filtering. The proposed approach is composed of two stages. In the first stage, multi-objective evolutionary feature selection is used to minimize both the misclassification cost of the proposed model and the number of attributes necessary for spam filtering. Then, the approach uses cost-sensitive ensemble learning techniques with regularized deep neural networks as base learners. We demonstrate that this approach is effective for social network spam filtering on two benchmark datasets. We also show that the proposed approach outperforms other popular algorithms used in social network spam filtering, such as random forest, Naïve Bayes or support vector machines. [ABSTRACT FROM AUTHOR]

Infertility affects one out of seven couples around the world. Therefore, the best possible management of the in vitro fertilization (IVF) treatment and patient advice is crucial for both patients and medical practitioners. The ultimate concern of the patients is the success of an IVF procedure, which depends on a number of influencing attributes. Without any automated tool, it is hard for the practitioners to assess any influencing trend of the attributes and factors that might lead to a successful IVF pregnancy. This paper proposes a hill climbing feature (attribute) selection algorithm coupled with automated classification using machine learning techniques with the aim to analyze and predict IVF pregnancy in greater accuracy. Using 25 attributes, we assessed the prediction ability of IVF pregnancy success for five different machine learning models, namely multilayer perceptron (MLP), support vector machines (SVM), C4.5, classification and regression trees (CART) and random forest (RF). The prediction ability was measured in terms of widely used performance metrics, namely accuracy rate, F-measure and AUC. Feature selection algorithm reduced the number of most influential attributes to nineteen for MLP, sixteen for RF, seventeen for SVM, twelve for C4.5 and eight for CART. Overall, the most influential attributes identified are: 'age', 'indication' of fertility factor, 'Antral Follicle Counts (AFC)', 'NbreM2', 'method of sperm collection', 'Chamotte', 'Fertilization rate in vitro', 'Follicles on day 14' and 'Embryo transfer day.' The machine learning models trained with the selected set of features significantly improved the prediction accuracy of IVF pregnancy success to a level considerably higher than those reported in the current literature. [ABSTRACT FROM AUTHOR]

In this paper, a hybrid system for wind power ramp events (WPREs) detection is proposed. The system is based on modeling the detection problem as a binary classification problem from atmospheric reanalysis data inputs. Specifically, a hybrid neuro-evolutionary algorithm is proposed, which combines artificial neural networks such as extreme learning machine (ELM), with evolutionary algorithms to optimize the trained models and carry out a feature selection on the input variables. The phenomenon under study occurs with a low probability, and for this reason the classification problem is quite unbalanced. Therefore, is necessary to resort to techniques focused on providing a balance in the classes, such as the synthetic minority over-sampling technique approach, the model applied in this work. The final model obtained is evaluated by a test set using both ELM and support vector machine algorithms, and its accuracy performance is analyzed. The proposed approach has been tested in a real problem of WPREs detection in three wind farms located in different areas of Spain, in order to see the spatial generalization of the method. [ABSTRACT FROM AUTHOR]

Methionine is a proteinogenic amino acid that can be post-translationally modified. It is now well established that reactive oxygen species can oxidise methionine residues within living cells. For a long time, it has been thought that such a modification represents merely an inevitable damage derived from aerobic metabolism. However, several authors have begun to contemplate a possible role for this methionine modification in cell signalling. During the last years, a number of proteomic studies have been carried out with the purpose of detecting proteins containing oxidised methionines. Although these proteomic works allow to pinpoint those methionines being oxidised, they are also arduous, expensive and time-consuming. For these reasons, computational approaches aimed at predicting methionine oxidation sites in proteins become an appealing alternative. In the current work, we address methionine oxidation prediction by combining computational intelligence methods with feature engineering and feature selection techniques to improve the efficacy of several machine learning models, while reducing the number of input characteristics needed to get high accuracy rates. We compare random forests, support vector machines, neural networks and flexible discriminant analysis models. Random forests give the best AUC ( 0.8124 ± 0.0334 ) and accuracy rates ( 0.7590 ± 0.0551 ) by using only a reduced set of 16 characteristics. These results surpass the outcomes of previous works. In addition, we present an end-user script that has been developed to take a protein ID as an input and return a list with the oxidation state of all the methionine residues found in the analysed protein. Finally, to illustrate the applicability of this tool, we have selected the human α 1 -antitrypsin protein as a case study. This protein was selected because it was not present among the set of proteins used to build up the predictive models but the protein has been well characterised experimentally in terms of methionine oxidation. The prediction returned by our script fully matches the empirical evidence. Out of the nine methionine residues found in this protein, our model predicts the oxidation of only two of them, M351 and M358, which have been reported, on the base of mass spectrometry analyses, to be particularly susceptible to oxidation. [ABSTRACT FROM AUTHOR]

Prostate cancer is the fourth most common cancer among all cancers and the second most common cancer in men. The rate of increase in prostate cancer incidence is higher than the overall increase of cancer incidents. 68% of prostate cancer cases are from developed countries. There has been very little research on the most suitable techniques for analysing prostate cancer gene expression datasets to identify those genes that may be most related to prostate cancer. This paper attempts to identify significant (influential) attributes in a well-established prostate cancer gene expression dataset consisting of over 12,533 attributes for 102 samples (50 normal, 52 tumour). Several (7) different statistical and artificial intelligence (AI)-based feature selection methods were paired with four different classifiers, namely ANNs, Naive Bayes, AdaBoost and J48. Prediction experiments are carried using ANNs with unseen sample testing. In our experiments, ANNs outperformed all other approaches for classification with sequential forward feature selection (SFFS), achieving 100% accuracy. Naive Bayes and AdaBoost achieved best accuracy of 96.3 and 93.13% with support vector machine (SVM) attribute selection, whereas J48 could get only 89.21% with SFFS approach. For prediction experiments, ANNs obtained an accuracy of 95.1% with SVM attribute selection (correctly predicting 96 out of 102 samples). Finally, by investigating National Center for Biotechnology Information database it is found that 21 out of 24 attributes (87.5%) that belong to SVM attribute selection have a reference to cancer/tumour, thereby establishing a link between feature selection and biological plausibility. The main contribution of this paper is in identifying the importance of pairing the most appropriate feature selection strategy with the most appropriate classification strategy when dealing with significantly underdetermined data. This paper also emphasizes differences and similarities between the influence of classification and prediction of prostate cancer. There is another new approach we considered while doing the classification and prediction experiments. Apart from using 7 different feature selection approaches, we have derived new set of attributes by adding all attributes (union), selecting common attributes (intersection) and rest of the attributes (not common). [ABSTRACT FROM AUTHOR]

In this paper, we propose a real-time online shopper behavior analysis system consisting of two modules which simultaneously predicts the visitor's shopping intent and Web site abandonment likelihood. In the first module, we predict the purchasing intention of the visitor using aggregated pageview data kept track during the visit along with some session and user information. The extracted features are fed to random forest (RF), support vector machines (SVMs), and multilayer perceptron (MLP) classifiers as input. We use oversampling and feature selection preprocessing steps to improve the performance and scalability of the classifiers. The results show that MLP that is calculated using resilient backpropagation algorithm with weight backtracking produces significantly higher accuracy and F1 Score than RF and SVM. Another finding is that although clickstream data obtained from the navigation path followed during the online visit convey important information about the purchasing intention of the visitor, combining them with session information-based features that possess unique information about the purchasing interest improves the success rate of the system. In the second module, using only sequential clickstream data, we train a long short-term memory-based recurrent neural network that generates a sigmoid output showing the probability estimate of visitor's intention to leave the site without finalizing the transaction in a prediction horizon. The modules are used together to determine the visitors which have purchasing intention but are likely to leave the site in the prediction horizon and take actions accordingly to improve the Web site abandonment and purchase conversion rates. Our findings support the feasibility of accurate and scalable purchasing intention prediction for virtual shopping environment using clickstream and session information data. [ABSTRACT FROM AUTHOR]

Support vector machines (SVM) are one of the important techniques used to solve classifications problems efficiently. Setting support vector machine kernel factors affects the classification performance. Feature selection is a powerful technique to solve dimensionality problems. In this paper, we optimized SVM factors and chose features using a Grasshopper Optimization Algorithm (GOA). GOA is a new heuristic optimization algorithm inspired by grasshoppers searching for food. It approved its ability to solve real-world problems with anonymous search space. We applied the proposed GOA + SVM approach on biomedical data sets for Iraqi cancer patients in 2010–2012 and for University of California Irvine data sets. [ABSTRACT FROM AUTHOR]

Support vector machine (SVM) is a well-regarded machine learning algorithm widely applied to classification tasks and regression problems. SVM was founded based on the statistical learning theory and structural risk minimization. Despite the high prediction rate of this technique in a wide range of real applications, the efficiency of SVM and its classification accuracy highly depends on the parameter setting as well as the subset feature selection. This work proposes a robust approach based on a recent nature-inspired metaheuristic called multi-verse optimizer (MVO) for selecting optimal features and optimizing the parameters of SVM simultaneously. In fact, the MVO algorithm is employed as a tuner to manipulate the main parameters of SVM and find the optimal set of features for this classifier. The proposed approach is implemented and tested on two different system architectures. MVO is benchmarked and compared with four classic and recent metaheuristic algorithms using ten binary and multi-class labeled datasets. Experimental results demonstrate that MVO can effectively reduce the number of features while maintaining a high prediction accuracy. [ABSTRACT FROM AUTHOR]

The aim of the paper is to automatically select the optimal EEG rhythm/channel combinations capable of classifying human alertness states. Four alertness states were considered, namely ‘engaged’, ‘calm’, ‘drowsy’ and ‘asleep’. The features used in the automatic selection are the energies associated with the conventional rhythms, δ,θ,α,β and γ, extracted from overlapping windows of the different EEG channels. The selection process consists of two stages. In the first stage, the optimal brain regions, represented by sets of EEG channels, are selected using a simple search technique based on support vector machine (SVM), extreme learning machine (ELM) and LDA classifiers. In the second stage, a fuzzy rule-based alertness classification system (FRBACS) is used to identify, from the previously selected EEG channels, the optimal features and their supports. The IF-THEN rules used in FRBACS are constructed using a novel differential evolution-based search algorithm particularly designed for this task. Each alertness state is represented by a set of IF-THEN rules whose antecedent parts contain EEG rhythm/channel combination. The selected spatio-frequency features were found to be good indicators of the different alertness states, as judged by the classification performance of the FRBACS that was found to be comparable to those of the SVM, ELM and LDA classifiers. Moreover, the proposed classification system has the advantage of revealing simple and easy to interpret decision rules associated with each of the alertness states. [ABSTRACT FROM AUTHOR]

In the past years, non-parallel plane classifiers that seek projection direction instead of hyperplane for each class have attracted much attention, such as the multi-weight vector projection support vector machine (MVSVM) and the projection twin support vector machine (PTSVM). Instead of solving two generalized eigenvalue problems in MVSVM, PTSVM solves two related SVM-type problems to obtain the two projection directions by solving two smaller quadratic programming problems, similar to twin support vector machine. In order to suppress input space features, we propose a novel non-parallel classifier to automatically select significant features, called regularization feature selection projection twin support vector machine (RFSPTSVM). In contrast to the PTSVM, we first incorporate a regularization term to ensure the optimization problems are convex, and then replace all the terms with L-norm ones. By minimizing an exterior penalty function of the linear programming problem and using a fast generalized Newton algorithm, our RFSPTSVM obtains very sparse solutions. For nonlinear case, the method utilizes minimal number of kernel functions. The experimental results on toy datasets, Myeloma dataset, several UCI benchmark datasets, and NDCC generated datasets show the feasibility and effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Upper body power (UBP) is a major parameter in cross-country skiing that strongly influences the performance of skiers during races. Although the direct measurement of UBP in laboratory environment provides the most accurate assessment, practical difficulties such as the need of sophisticated and expensive testing equipment and trained stuff have given rise to the necessity of predicting and identifying the relevant predictors of UBP. This study aims to develop feature selection-based prediction models to predict 10-s UBP (UBP) and 60-s UBP (UBP) of cross-country skiers with the help of support vector machines (SVMs) using different kernel functions. For model testing, tenfold cross-validation has been carried out, and the performance of the models has been assessed by computing their standard error of estimates (SEEs) and multiple correlation coefficients ( Rs). With respect to the obtained results, the lowest SEEs (27.81 and 18.73 W) and the highest Rs (0.92 and 0.94) for the most accurate UBP and UBP prediction models have been obtained by SVM using the radial basis function kernel (SVM-RBF), respectively. The predictor variables gender and age have been found to be the most important variables in predicting the UBP and UBP of cross-country skiers. Also, the results have been compared to those of multilayer perceptron and multiple linear regression-based prediction models. Prediction models based on SVM-RBF exhibit much lower SEEs and higher Rs than the prediction models developed by other regression methods and can be safely utilized in making UBP predictions within acceptable limits of accuracy. [ABSTRACT FROM AUTHOR]

In this paper, we propose novel methods to find the best relevant feature subset using fuzzy rough set-based attribute subset selection with biologically inspired algorithm search such as ant colony and particle swarm optimization and the principles of an evolutionary process. We then propose a hybrid fuzzy rough with K-nearest neighbor (K-NN)-based classifier (FRNN) to classify the patterns in the reduced datasets, obtained from the fuzzy rough bio-inspired algorithm search. While exploring other possible hybrid evolutionary processes, we then conducted experiments considering (i) same feature selection algorithm with support vector machine (SVM) and random forest (RF) classifier; (ii) instance based selection using synthetic minority over-sampling technique with fuzzy rough K-nearest neighbor (K-NN), SVM and RF classifier. The proposed hybrid is subsequently validated using real-life datasets obtained from the University of California, Irvine machine learning repository. Simulation results demonstrate that the proposed hybrid produces good classification accuracy. Finally, parametric and nonparametric statistical tests of significance are carried out to observe consistency of the classifiers. [ABSTRACT FROM AUTHOR]

Parameter optimization and feature selection influence the classification accuracy of support vector machine (SVM) significantly. In order to improve classification accuracy of SVM, this paper hybridizes chaotic search and gravitational search algorithm (GSA) with SVM and presents a new chaos embedded GSA-SVM (CGSA-SVM) hybrid system. In this system, input feature subsets and the SVM parameters are optimized simultaneously, while GSA is used to optimize the parameters of SVM and chaotic search is embedded in the searching iterations of GSA to optimize the feature subsets. Fourteen UCI datasets are employed to calculate the classification accuracy rate in order to evaluate the developed CGSA-SVM approach. The developed approach is compared with grid search and some other hybrid systems such as GA-SVM, PSO-SVM and GSA-SVM. The results show that the proposed approach achieves high classification accuracy and efficiency compared with well-known similar classifier systems. [ABSTRACT FROM AUTHOR]

In this report, the parameters identification of a proportional-integral-derivative (PID) algorithm implemented in a programmable logic controller (PLC) using support vector regression (SVR) is presented. This report focuses on a black box model of the PID with additional functions and modifications provided by the manufacturers and without information on the exact structure. The process of feature selection and its impact on the training and testing abilities are emphasized. The method was tested on a real PLC (Siemens and General Electric) with the implemented PID. The results show that the SVR maps the function of the PID algorithms and the modifications introduced by the manufacturer of the PLC with high accuracy. With this approach, the simulation results can be directly used to tune the PID algorithms in the PLC. The method is sufficiently universal in that it can be applied to any PI or PID algorithm implemented in the PLC with additional functions and modifications that were previously considered to be trade secrets. This method can also be an alternative for engineers who need to tune the PID and do not have any such information on the structure and cannot use the default settings for the known structures. [ABSTRACT FROM AUTHOR]

In a CAD system for the detection of masses, segmentation of mammograms yields regions of interest (ROIs), which are not only true masses but also suspicious normal tissues that result in false positives. We introduce a new method for false-positive reduction in this paper. The key idea of our approach is to exploit the textural properties of mammograms and for texture description, to use Weber law descriptor (WLD), which outperforms state-of-the-art best texture descriptors. The basic WLD is a holistic descriptor by its construction because it integrates the local information content into a single histogram, which does not take into account the spatial locality of micropatterns. We extend it into a multiscale spatial WLD (MSWLD) that better characterizes the texture micro structures of masses by incorporating the spatial locality and scale of microstructures. The dimension of the feature space generated by MSWLD becomes high; it is reduced by selecting features based on their significance. Finally, support vector machines are employed to classify ROIs as true masses or normal parenchyma. The proposed approach is evaluated using 1024 ROIs taken from digital database for screening mammography and an accuracy of Az = 0.99 ± 0.003 (area under receiver operating characteristic curve) is obtained. A comparison reveals that the proposed method has significant improvement over the state-of-the-art best methods for false-positive reduction problem. [ABSTRACT FROM AUTHOR]

This paper presents a new version of support vector machine (SVM) named l2 −  lp SVM (0 <  p < 1) which introduces the lp-norm (0 <  p < 1) of the normal vector of the decision plane in the standard linear SVM. To solve the nonconvex optimization problem in our model, an efficient algorithm is proposed using the constrained concave–convex procedure. Experiments with artificial data and real data demonstrate that our method is more effective than some popular methods in selecting relevant features and improving classification accuracy. [ABSTRACT FROM AUTHOR]

Stock index forecasting is one of the most difficult tasks that financial organizations, firms and private investors have to face. Support vector regression (SVR) has become a popular alternative in stock index forecasting tasks due to its generalization capability in obtaining a unique solution. However, the major limitation of SVR is that it cannot capture the relative importance of independent variables to the dependent variable when many potential independent variables are considered. This study incorporates feature selection method and SVR for building stock index forecasting model. The proposed model uses multivariate adaptive regression splines (MARS), an effective nonlinear and nonparametric regression methodology, to identify important forecasting variables. The obtained significant predictor variables are then served as the inputs for the SVR model. Experimental results reveal that the obtained important variables from MARS can improve the forecasting performance of the SVR models. Moreover, the MARS results provide useful information about the relationship between the selected predictor variables and stock index through the obtained basis functions, important predictor variables and the MARS prediction function. Hence, the proposed stock index forecasting model can generate good forecasting performance and exhibits the capability of identifying significant predictor variables, which provide valuable information for further investment decisions/strategies. [ABSTRACT FROM AUTHOR]

This paper focuses on feature selection in classification. A new version of support vector machine (SVM) named p-norm support vector machine ( $$p\in[0,1]$$) is proposed. Different from the standard SVM, the p-norm $$(p\in[0,1])$$ of the normal vector of the decision plane is used which leads to more sparse solution. Our new model can not only select less features but also improve the classification accuracy by adjusting the parameter p. The numerical experiments results show that our p-norm SVM is more effective than some usual methods in feature selection. [ABSTRACT FROM AUTHOR]

The speech signal consists of linguistic information and also paralinguistic one such as emotion. The modern automatic speech recognition systems have achieved high performance in neutral style speech recognition, but they cannot maintain their high recognition rate for spontaneous speech. So, emotion recognition is an important step toward emotional speech recognition. The accuracy of an emotion recognition system is dependent on different factors such as the type and number of emotional states and selected features, and also the type of classifier. In this paper, a modular neural-support vector machine (SVM) classifier is proposed, and its performance in emotion recognition is compared to Gaussian mixture model, multi-layer perceptron neural network, and C5.0-based classifiers. The most efficient features are also selected by using the analysis of variations method. It is noted that the proposed modular scheme is achieved through a comparative study of different features and characteristics of an individual emotional state with the aim of improving the recognition performance. Empirical results show that even by discarding 22% of features, the average emotion recognition accuracy can be improved by 2.2%. Also, the proposed modular neural-SVM classifier improves the recognition accuracy at least by 8% as compared to the simulated monolithic classifiers. [ABSTRACT FROM AUTHOR]

This paper proposes a new method for fuzzy rule extraction from trained support vector machines (SVMs) for multi-class problems, named FREx_SVM. SVMs have been used in a variety of applications. However, they are considered 'black box models,' where no interpretation about the input-output mapping is provided. Some methods to reduce this limitation have already been proposed, but they are restricted to binary classification problems and to the extraction of symbolic rules with intervals or functions in their antecedents. In order to improve the interpretability of the generated rules, this paper presents a new model for extracting fuzzy rules from a trained SVM. The proposed model is suited for classification in multi-class problems and includes a wrapper feature selection algorithm. It is evaluated in four benchmark databases, and results obtained demonstrate its capacity to generate a reduced set of interpretable fuzzy rules that explains both the classification database and the influence of each input variable on the determination of the final class. [ABSTRACT FROM AUTHOR]

Auditory brainstem response (ABR) has become a routine clinical tool in neurological and audiological assessment. ABR measurement process with ensemble averaging is very time-consuming and uncomfortable for subjects due to the more repetition of single trials. This condition also restricts the wide usability of ABR in clinical applications. Therefore, the reduction in repetitions has a great importance in ABR measurements. In this study, 488 ABR responses are used for creating two different data sets. The first set is created conventionally by ensemble averaging of 1,024 single trials for each ABR pattern. The second set is obtained from the first estimated 64 single trials of the same records for each ABRs. Estimation is realized by using a nonlinear adaptive filtering algorithm. In classification stage, a powerful classifier integrated with a feature selection algorithm is performed for each data set. In result, the classification performance for estimated ABR data with 64 repetitions is better than the classification performance of the ensemble averaged data with 1,024 repetitions. The proposed system is resulted in an accuracy of 96% for estimated ABRs. So, the proposed system can effectively be used for threshold detection in auditory assessment providing a high accuracy. While the obtained results contribute to the practical ABR usage in clinics, the great significance of it arises from the reduction in repetitions via estimation of ABRs. [ABSTRACT FROM AUTHOR]

The present paper investigates the application of the multi-layer perceptron (MLP) to the task of categorizing texts based on their authors’ style. This task is of particular importance for information retrieval applications involving very large document databases. The emphasis of this article is to determine the extent to which the MLP model can be fine-tuned to successfully analyse such data, uncovering the stylistic differences among authors. The MLP-based method is compared and contrasted to statistical techniques, such as discriminant analysis, that are widely used in stylistic studies. The comparison of the methods is based on their classification performance, to provide an objective evaluation of the advantages of each method. A second aim of the study presented here is to compare the effectiveness of distinct features in the task of uncovering the author identity for each method. To evaluate to a greater depth the effectiveness of the entire approach, the results of the proposed MLP-based method are compared to those of established approaches, such as the support vector machines (SVM), using both the original parameters employed by the MLP as well as term frequency–inverse document frequency (TF–IDF) parameters, and the cascade correlation approach. It is found that the proposed MLP-based approach possesses a number of advantages, such as high classification accuracy, broadly comparable to that of the SVM, coupled with the ability to algorithmically reduce the set of parameters used without adversely affecting the classification accuracy. [ABSTRACT FROM AUTHOR]

Multivariate calibration is a classic problem in the analytical chemistry field and frequently solved by partial least squares (PLS) and artificial neural networks (ANNs) in the previous works. The spaciality of multivariate calibration is high dimensionality with small sample. Here, we apply support vector regression (SVR) as well as ANNs, and PLS to the multivariate calibration problem in the determination of the three aromatic amino acids ( phenylalanine, tyrosine and tryptophan) in their mixtures by fluorescence spectroscopy. The results of the leave-one-out method show that SVR performs better than other methods, and appear to be one good method for this task. Furthermore, feature selection is performed for SVR to remove redundant features and a novel algorithm named Prediction RIsk based FEature selection for support vector Regression (PRIFER) is proposed. Results on the above multivariate calibration data set show that PRIFER is a powerful tool for solving the multivariate calibration problems. [ABSTRACT FROM AUTHOR]