Different approaches have been proposed for FEATURE selection to obtain suitable FEATUREs subset among all FEATUREs. These methods search FEATURE space for FEATURE subsets which satisfies some criteria or optimizes several objective functions. The objective functions are divided into two main groups: filter and wrapper methods. In filter methods, FEATUREs subsets are selected due to some measures like inter-class distance, FEATUREs statistical independence or information theoretic measures. Even though, wrapper methods use a classifier to evaluate FEATUREs subsets by their predictive accuracy (on test data) by statistical resampling or cross-validation. Filter methods usually use only one measure for FEATURE selection that does not necessarily produce the best result. In this paper, we proposed to use the classification error measures besides to filter measures where our classifier is support VECTOR machine (SVM). To this end, we use multi objective genetic algorithm. In this way, one of our FEATURE selection measure is SVM classification error. Another measure is selected between mutual information and Laplacian criteria which indicates informative content and structure preserving property of FEATUREs, respectively. The evaluation results on the UCI datasets show the efficiency of this method.

Unsupervised methods have become essential for managing and analyzing large datasets due to their ability to uncover underlying patterns without labeled data. In particular, unsupervised learning techniques have shown promise in enhancing the robustness of FEATURE VECTORs, which can subsequently improve the accuracy and efficiency of classification models. This study presents a novel approach called Automatic Weighted Clustering (AWC), designed as an unsupervised algorithm specifically aimed at optimizing data segmentation to support and enhance the classification performance of supervised algorithms. The AWC algorithm, applied in the context of medical datasets, enables efficient knowledge discovery by automatically identifying and weighting important data FEATUREs. To assess AWC’s performance, we conducted extensive experiments across various datasets to test the algorithm’s generalizability and scalability. The AWC approach yielded nine distinct clusters from the dataset, demonstrating a 6.26% accuracy improvement compared to the standard FEATURE VECTOR and a 4.83% increase over the traditional K-means clustering method. Additionally, AWC achieved a 30.68% reduction in execution time, highlighting its potential for faster, more accurate medical data analysis and classification.

Electromyographs are used for electromyography signal extraction from neurologically activated muscle cells. These signals are investigated to extract discriminating patterns to be categorized in the classification stage of myoelectric control systems (MCSs) designed for various applications. FEATURE extraction is a fundamental step in EMG signal processing which affects the overall performance of MCSs. To improve classification accuracy of MCSs, this paper proposes a novel approach for FEATURE extraction from time-frequency images of EMG signals using local binary patterns and gray level co-occurrence matrices. In contrast to time alone and frequency alone approaches, by textural analysis of EMG signal spectrogram, time-frequency patterns of these signals are revealed, simultaneously. Furthermore, LBP and GLCM expose relational properties of time-frequency patterns which areexploited as the main FEATUREs for classification. EMG physical action dataset is utilized in this study to evaluate the proposed method. In the classification stage, support VECTOR machine classifiers are used in two segmented and holistic modes. The best classification accuracy of 98. 75% is obtained by segmented approach which is superior to the results provided by state of the art methods.

Breast cancer is known to be among the most prevalent cause of mortality among women. Since early breast cancer diagnosis increases survival chances, the development of a system with a highly accurate output to detect suspicious masses in mammographic images is of great significance. Thus, many studies have focused on the development of methods with favorable performance and acceptable accuracy to detect cancerous masses, proposed various techniques to diagnose breast cancer, and compared their accuracies. Most previous studies have used composite selection and FEATURE reduction techniques to detect breast cancer and accelerate its treatment; however, most have failed to reach the desired accuracy due to the selection of ineffective FEATUREs and the lack of a proper analytical method for the FEATUREs. The present study reviews the methods proposed to detect breast cancer so far and analyzes the process of FEATURE VECTOR optimization techniques as well as the normal/abnormal and benign/malignant mass classification.

Recently, a large set of electroencephalography (EEG) data is being generated by several high-quality labs worldwide and is free to be used by all researchers in the world. On the other hand, many neuroscience researchers need these data to study different neural disorders for better diagnosis and evaluating the treatment. However, some format adaptation and pre-processing are necessary before using these available data. In this paper, we introduce the SecondBrain as a new lightweight and simplified module that can easily apply various major analysis on EEG data with common data formats. The characteristics of the SecondBrain shows that it is suitable for everyday usage with medium analyzing power. It is easy to learn and accept many data formats. The SecondBrain module has been developed with Python and has the power to windowing data, whitening transform, independent component analysis (ICA), downloading the public datasets, computing common spatial patterns (CSP) and other useful analysis. The SecondBrain, also, employs a common spatial pattern (CSP) to extract FEATUREs and classifying the EEG MI-based data through support VECTOR machine (SVM). We achieved a satisfactory result in terms of speed and performance.

Inertial Navigation System (INS) is one of the navigation systems widely used in various land-based, aerial, and marine applications. Among all types of INS, Microelectromechanical System (MEMS)-based INS can be widely utilized, owing to their low cost, lightweight, and small size. However, due to the manufacturing technology, MEMS-based INS suffers from deterministic and stochastic errors, which increase positioning errors over time. In this paper, a new effective noise reduction method is proposed that can provide more accurate outputs of MEMS-based inertial sensors. The intelligent method in this paper is a combined denoising method that combines Wavelet Transform (WT), Permutation Entropy (PE), Support VECTOR Regression (SVR), and Genetic Algorithm (GA). Firstly, WT is employed to obtain a time-frequency representation of raw data. Secondly, a four-element FEATURE VECTOR is formed. These four FEATUREs are (1) amplitude of frequency, (2) its ratio to mean of amplitudes of all frequencies, (3) location of frequency in time-frequency representation, and (4) judgment on behaviors of frequency that is obtained by utilizing PE. Thirdly, based on the FEATURE VECTOR, the GA-SVR algorithm predicts amplitudes of all frequencies in the time-frequency representation of the denoised signal. Finally, by employing inverse WT the denoised signal is obtained. In this work, the outputs of the Inertial Measurement Unit (IMU) in ADIS16407 sensor, as a low-cost and MEMS-based INS, have been utilized for data collection. The proposed method has been compared with other noise reduction methods and the achieved results verify superior improvement than other methods.