Genetic Based Feature Selection Research Articles

Improving forest above-ground biomass estimation using genetic-based feature selection from Sentinel-1 and Sentinel-2 data (case study of the Noor forest area in Iran)

Biomass holds great importance in the environment, as it not only allows us to measure the carbon stored in forests but also facilitates the assessment of biodiversity and the evaluation of ecological integrity within these crucial ecosystems. In this study, we employed a Genetic Algorithm (GA) to estimate forest Above-Ground Biomass (AGB) by selecting the most applicable features from both Sentinel-2 optical and Sentinel-1 Synthetic Aperture Radar (SAR) images in the Noor forest. The study area was divided into four distinct regions (north, near north, middle, and south), and each region was documented with 100 sample plots through fieldwork to enable comprehensive analysis. In our workflow, Sentinel-2-derived features (i.e., spectral bands, vegetation indices (VIs), soil indices (SIs), and water indices (WIs), along with Sentinel-1 SAR features were initially extracted. Subsequently, GA was employed to select the most optimal features among them within both Random Forest (RF) and Multiple Linear Regression (MLR) models, leading to enhanced accuracy in the forest AGB estimation process. The experimental results demonstrated that the RF model outperformed the MLR model in estimating forest AGB. Furthermore, incorporating GA-based feature selection substantially improved the accuracy of both models, resulting in more dependable AGB estimations. The selected features from the combined Sentinel-1 and Sentinel-2 data also provided the best AGB estimation, surpassing the individual use of each dataset. The selected features from Sentinel-2 particularly played a more substantial role in achieving this overall enhanced performance in AGB estimation. The AGB estimates based on GA-RF were more accurate in all cases, with an average coefficient of determination (R2) of 0.5 and average RMSE of 13.17 Mg ha−1, while the MLR-based estimates were less accurate, with an average R2 value lower than 0.3 and average RMSE higher than 16 Mg ha−1. Furthermore, the GA-RF model selected a wider variety of features including spectral bands, indices, and SAR features compared to GA-MLR, resulting in accurate AGB estimation in the Noor forest.

RNN-LSTM Based Deep Learning Model for Tor Traffic Classification

ABSTRACT Tor is an anonymous browser software running on an overlay network. Due to the nature of the end-to-end encryption channel, it is hard to analyse the network traffic. Thus, intruders prefer the Tor browser to hide their identity and access the offensive content. Tor relays are secure from network monitoring, tracking and surveillance. There are so many research contributions for tracking the network traffic and classifying it based on various features and attributes. In this paper, we explained RNN-LSTM-based deep learning model to classify the network traffic based on their nature Tor/non-Tor. We have tested the model with open data sets ISCXTor2016 data sets and samples retrieved in our environment using CIC-flowmeter-4.0. The binary classification model using RNN-LSTM classifies the network traffic with better accuracy and precision. The same experiment conducted in the traditional deep neural network model provides large false positives and false negatives. Here we also present a detailed study and analysis of the model compare with ANN classifiers and genetic-based feature selection method.

Genetic-based feature selection for efficient motion imaging of a brain–computer interface framework

Objective. A brain–computer interface (BCI) equips humans with the ability to control computers and technical devices mentally. However, the enormous data and the existing irrelevant features of the electrocorticogram signal limit the performance of the classifier. To address these problems, a novel signal processing framework for a binary motor imagery-based BCI system (MI–BCI) is proposed in this paper. Approach. Stockwell transform and Bayesian linear discriminant analysis were applied to feature extraction and classification, respectively, and a genetic algorithm (GA) was used in the process of feature selection to extract the most relevant features for classification. The superiority of the algorithm is demonstrated through test results based on the BCI Competition III dataset I. Main results. By comparing the processes with or without feature selection, the performance of the classification was proven to improve using the GA. By adjusting the parameters of the GA, the best feature set (selected 48.6% features) was selected to achieve classification sensitivity, specificity, precision, and accuracy of 94%, 98%, 97.9%, and 96%, respectively, exceeding the results of the existing state-of-the art algorithms. Significance. As the proposed method can reduce the number of features and select the best feature set, its classification performance was improved and the classification time was shortened; thus, it can be applied to various BCI systems.

An effective and efficient approach to classification with incomplete data

Many real-world datasets suffer from the unavoidable issue of missing values. Classification with incomplete data has to be carefully handled because inadequate treatment of missing values will cause large classification errors. Using imputation to transform incomplete data into complete data is a common approach to classification with incomplete data. However, simple imputation methods are often not accurate, and powerful imputation methods are usually computationally intensive. A recent approach to handling incomplete data constructs an ensemble of classifiers, each tailored to a known pattern of missing data. The main advantage of this approach is that it can classify new incomplete instances without requiring any imputation. This paper proposes an improvement on the ensemble approach by integrating imputation and genetic-based feature selection. The imputation creates higher quality training data. The feature selection reduces the number of missing patterns which increases the speed of classification, and greatly increases the fraction of new instances that can be classified by the ensemble. The results of experiments show that the proposed method is more accurate, and faster than previous common methods for classification with incomplete data.

Image-based emotion recognition using evolutionary algorithms

Abstract In pattern recognition, the classification accuracy has a strong correlation with the selected features. Therefore, in the present paper, we applied an evolutionary algorithm in combination with linear discriminant analysis (LDA) to enhance the feature selection in a static image-based facial expressions system. The accuracy of the classification depends on whether the features are well representing the expression or not. Therefore the optimization of the selected features will automatically improve the classification accuracy. The proposed method not only improves the classification but also reduces the dimensionality of features. Our approach outperforms linear-based dimensionality reduction algorithms and other existing genetic-based feature selection algorithms. Further, we compare our approach with VGG (Visual Geometry Group)-face convolutional neural network (CNN), according to the experimental results, the overall accuracy is 98.67% for either our approach or VGG-face. However, the proposed method outperforms CNN in terms of training time and features size. The proposed method proves that it is able to achieve high accuracy by using far fewer features than CNN and within a reasonable training time.

Improved occupancy monitoring in non-domestic buildings

Measuring occupancy can facilitate energy efficiency in non-domestic buildings, when control systems are able to adjust heating and cooling based on demand rather than fixed schedules. The variable “occupancy profile” itself is rarely considered as a control system parameter in building energy management systems (BEMS), and this is largely because reliably measuring occupancy in the past has been too difficult, expensive, or a mixture of both. Occupancy detection is possible using e.g. CO2 sensors, passive infra-red (PIR) detectors, which can provide a basic trigger for services, but the actual occupancy count, and therefore the expected load on building services, requires a step change in instrumentation. Advanced occupancy sensors developed from a heterogeneous multisensory fusion strategy offer this, improving control system performance, e.g. turning off services out of hours, and not over-ventilating, saving energy, while not under-ventilating during occupancy, benefitting comfort and health. While this is the case, there is a shortage of any systematic methodology for developing robust and reliable occupancy monitoring systems from heterogeneous multi-sensory sources. In this paper we describe an innovative sensor fusion approach utilising symmetrical uncertainty (SU) analysis and a genetic based feature selection for building occupancy estimation.

A Genetic-Based Feature Selection Approach in the Identification of Left/Right Hand Motor Imagery for a Brain-Computer Interface.

Electroencephalography is a non-invasive measure of the brain electrical activity generated by millions of neurons. Feature extraction in electroencephalography analysis is a core issue that may lead to accurate brain mental state classification. This paper presents a new feature selection method that improves left/right hand movement identification of a motor imagery brain-computer interface, based on genetic algorithms and artificial neural networks used as classifiers. Raw electroencephalography signals are first preprocessed using appropriate filtering. Feature extraction is carried out afterwards, based on spectral and temporal signal components, and thus a feature vector is constructed. As various features might be inaccurate and mislead the classifier, thus degrading the overall system performance, the proposed approach identifies a subset of features from a large feature space, such that the classifier error rate is reduced. Experimental results show that the proposed method is able to reduce the number of features to as low as 0.5% (i.e., the number of ignored features can reach 99.5%) while improving the accuracy, sensitivity, specificity, and precision of the classifier.

Spam filtering by using Genetic based Feature Selection

Spam is defined as redundant and unwanted electronical letters, and nowadays, it has created many problems in business life such as occupying networks bandwidth and the space of user’s mailbox. Due to these problems, much research has been carried out in this regard by using classification technique. The resent research show that feature selection can have positive effect on the efficiency of machine learning algorithm. Most algorithms try to present a data model depending on certain detection of small set of features. Unrelated features in the process of making model result in weak estimation and more computations. In this research it has been tried to evaluate spam detection in legal electronica letters, and their effect on several Machin learning algorithms through presenting a feature selection method based on genetic algorithm. Bayesian network and KNN classifiers have been taken into account in classification phase and spam base dataset is used.

Feature Subset Selection Based on the Genetic Algorithm

This paper presents a genetic-based feature selection algorithm for object recognition. Firstly, the proposed algorithm encodes a solution with a binary chromosome. Secondly, the initial population was generated randomly. Thirdly, a crossover operator and a mutation operator are employed to operate on these chromosomes to generate more competency chromosomes. The probability of the crossover and mutation are adjusted dynamically according to the generation number and the fitness value. The proposed algorithm is tested using the features extracted from cotton foreign fiber objects. The results indicate that the proposed algorithm can obtain the optimal feature subset, and can reduce the classification time while keeping the classification accuracy constant.