A Feature Extraction and Selection Framework for Electrocorticography-Based Neural Activity Classification.

Emotion recognition based on EEG signals is a critical area within affective computing, with applications in mental health monitoring, human-computer interaction, and neuroadaptive systems. However, accurately classifying emotional states from inherently non-stationary and noisy EEG data remains a major challenge. This study explores the classification of three discrete emotions, Humorous, Sad, and Fearful, elicited through video stimuli, using EEG recordings from six participants acquired via a 19-channel Mitsar amplifier at a 500 Hz sampling rate. Preprocessing steps included bandpass filtering (1–40 Hz), epoch segmentation, and multi-domain feature extraction encompassing statistical measures, spectral features, differential entropy, Hjorth parameters, and hemispheric asymmetry indicators. Data augmentation was applied to balance class distributions, particularly for the underrepresented fear category. The resulting features were normalized and structured to support temporal deep learning and classical machine learning models. The classification performance of Long Short-Term Memory (LSTM) networks was evaluated alongside Support Vector Machine (SVM), k-Nearest Neighbors (k-NN), and Random Forest (RF) classifiers. While LSTM demonstrated competency in capturing temporal dependencies, especially in fear recognition, SVM achieved the highest overall accuracy, 94.12%, outperforming LSTM at 85.16%, RF at 90.00%, and k-NN at 78.01%. These results suggest that when robust and discriminative features are employed, traditional models like SVM can surpass deep learning methods, particularly in small-scale EEG datasets with limited temporal complexity. This study underscores the importance of aligning model architecture with feature representation and contributes a comparative evaluation framework for EEG-based emotion recognition systems.

Platform X is widely used by students to express emotions related to academic activities, yet it has not been optimally analyzed. This study aims to classify student emotions and compare the performance of Support Vector Machine (SVM) and Long Short-Term Memory (LSTM). The method used is a computational experiment involving data crawling, preprocessing, lexicon-based labeling, and modeling using TF-IDF for SVM and word embedding for LSTM. The dataset consists of 2,914 data points. The results show that SVM achieved an accuracy of 71.18%, outperforming LSTM at 68.95%. This indicates that SVM is more stable on limited datasets, while LSTM tends to overfit. The study concludes that algorithm selection should consider data characteristics and contributes to the comparison of machine learning and deep learning methods. Keywords: LSTM, Sentiment Analysis, Social Media X, Student Emotion, SVM, TF-IDF

Industrial robots are widely used in critical tasks such as assembly, welding, and material handling as core components of modern manufacturing systems. For the reliable operation of these systems, early and accurate detection of execution failures is crucial. In this study, a comprehensive comparison of machine learning and deep learning methods is conducted for the classification of robot execution failures using data acquired from force-torque sensors. Three different feature engineering approaches are proposed. The first is a Baseline approach that includes 90 raw time-series features. The second is the Domain-6 approach, which consists of 6 basic statistical features per sensor (36 in total). The third is the Domain-12 approach, which comprises 12 comprehensive statistical features per sensor (72 in total). The domain features include the mean, standard deviation, minimum, maximum, range, slope, median, skewness, kurtosis, RMS, energy, and IQR. In total, ten classification algorithms are evaluated, including eight machine learning methods and two deep learning models: Support Vector Machines (SVM), Random Forest (RF), k-Nearest Neighbors (KNN), Artificial Neural Network (ANN), Naive Bayes (NB), Decision Trees (DT), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM-LGBM), as well as a One-Dimensional Convolutional Neural Network (CNN-1D) and Long Short-Term Memory (LSTM). For traditional machine learning algorithms, 5 × 5 nested cross-validation is used, whereas for deep learning models, 5-fold cross-validation with a 20% validation split is employed. To ensure statistical reliability, all experiments are repeated over 30 independent runs. The experimental results demonstrate that feature engineering has a decisive impact on classification performance. In addition, regardless of the feature set, the highest accuracy (93.85% ± 0.90) is achieved by the Naive Bayes classifier using the Baseline features. The Domain-12 feature set provides consistent improvements across many algorithms, with substantial performance gains. The results are reported using accuracy, precision, recall, and F1-score metrics and are supported by confusion matrices. Finally, permutation feature importance analysis indicates that the skewness features of the Fx and Fy sensors are the most critical variables for failure detection. Overall, these findings show that time-domain statistical features offer an effective approach for robot failure classification.

The hybrid Brain-Computer Interface [BCI] system gives an insight on the development of useful interfaces for users with different backgrounds, from medical applications to video games, where standalone and wearable means accessibility for the user. Systems such as EPOC offers a simple solution for acquiring electroencephalography and electromyography signals with low price and fast setup, compared to high tech medical equipment. From the processing point of view, a computer always offers the main foundation for solving any issue, as the Raspberry Pi [RPi] does, which provides the sufficient computational power for a BCI to be implemented and an open source operating system such as Raspbian. Certainly a wireless communication is a must between the robot and the RPi, where an Xbee module gives a simple bidirectional connection. Python is the principal tool used in the project with multiple libraries for the processing of brain and muscular signals not only for the preparation of them but classification as well, from multithreading functions, feature extraction such as power spectral density and Hjorth parameters, and a support vector machine classifiera.

The Army Rapid Capabilities Office (RCO) sponsored a Blind Signal Classification Competition seeking algorithms to automatically identify the modulation schemes of RF signal from complex-valued IQ (in-phase quadrature) samples. Traditional spectrum sensing technology uses energy detection to detect the existence of RF signals but the RCO competition further aimed to detect the modulation scheme of signals without prior information. Machine Learning (ML) technologies have been widely used for blind signal classification problem. Traditional ML methods usually have two stages where the first stage is to manually extract the features of the IQ symbols by subject matter experts and the second stage is to feed the features to an ML algorithm (e.g., a support vector machine) to develop the classifier. The state-of-art technology is to apply deep learning technologies such as Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN) directly to the complex-value IQ symbols to train a multi-class classifier. Our team, dubbed Deep Dreamers, participated in the RCO competition and placed 3rd out of 42 active teams across industry, academia, and government. In this work we share our experience and lessons learned from the competition. Deep learning methods such as CNN, Residual Neural Network (ResNet), and Long Short-Term Memory (LSTM) are the fundamental neural network layers we used to develop a multi-class classifier. None of our individual models were able to achieve a competitively high ranking in the competition. The key to our success was to use ensemble learning to average the outputs of multiple diverse classifiers. In order for ensemble methods to be more accurate than any of its base models; the base learners have to be as accurate as possible. We found that while ResNet was more accurate than the LSTM; the LSTM was less sensitive to deviations in the test set.

Automating real-time sign language translation through deep learning and machine learning techniques can greatly enhance communication between the deaf community and the wider public. This research investigates how these technologies can change the way individuals with speech impairments communicate. Despite advancements, developing accurate models for recognizing both static and dynamic gestures remains challenging due to variations in gesture speed and length, which affect the effectiveness of the models. We introduce a hybrid approach that merges machine learning and deep learning methods for sign language recognition. We provide new model for the recognition of Kazakh Sign Language (QazSL), employing five algorithms: Support Vector Machine (SVM), Long Short Term Memory (LSTM), Gated Recurrent Unit (GRU), Convolutional Neural Networks (CNN) with VGG19, ResNet-50, and YOLOv5. The models were trained on a QazSL dataset of more than 4,400 photos. Among the assessed models, the GRU attained the highest accuracy of 100%, followed closely by SVM and YOLOv5 at 99.98%, VGG19 at 98.87% for dynamic dactyls, LSTM at 85%, and ResNet-50 at 78.61%. These findings illustrate the comparative efficacy of each method in real-time gesture recognition. The results yield significant insights for enhancing sign language recognition systems, presenting possible advancements in accessibility and communication for those with hearing impairments.

Integrated phenology and climate in rice yields prediction using machine learning methods

Recently, there has been a remarkable growth in Artificial Intelligence (AI) with the development of efficient AI models and high-power computational resources for processing complex datasets. There has been a growing number of applications of machine learning in satellite remote sensing image data processing. In this work, machine learning methods were applied for crop classification of temporal multi- spectral satellite image to achieve better prediction of crop-wise area statistics. In India, agriculture has a huge impact on the national economy and most of the critical decisions are dependent on agricultural statistics. Sentinel-2 satellite image data for the Guntur district region of Andhra Pradesh, India has been used as the study area. The main reason for selecting this region is the diversity of agricultural crops and the availability of ground truth. As a baseline, the performance of machine learning models like Support Vector Machine (SVM), Random Forest (RF), Decision Tree (DT), and K-Nearest Neighbor (KNN) on crop classification was evaluated on 2016 single time multi-spectral satellite image. SVM performed well with an F1 score of 0.94. Further, the performance of SVM, RF, 1D and 2D Convolutional Neural Network (CNN), Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) and RNN with Gated Recurrent Unit (GRU) on 2018 Kharif season temporal multi-spectral satellite image has been evaluated for estimation of crop areas. The results show that SVM has the best F1 score of 0.99 and achieved a 95.9% agreement with the ground surveyed crop areas.

In this modern world, we are accustomed to a constant stream of data. Major social media sites like Twitter, Facebook, or Quora face a huge dilemma as a lot of these sites fall victim to spam accounts. These accounts are made to trap unsuspecting genuine users by making them click on malicious links or keep posting redundant posts by using bots. This can greatly impact the experiences that users have on these sites. A lot of time and research has gone into effective ways to detect these forms of spam. Performing sentiment analysis on these posts can help us in solving this problem effectively. The main purpose of this proposed work is to develop a system that can determine whether a tweet is “spam” or “ham” and evaluate the emotion of the tweet. The extracted features after preprocessing the tweets are classified using various classifiers, namely, decision tree, logistic regression, multinomial naïve Bayes, support vector machine, random forest, and Bernoulli naïve Bayes for spam detection. The stochastic gradient descent, support vector machine, logistic regression, random forest, naïve Bayes, and deep learning methods, namely, simple recurrent neural network (RNN) model, long short-term memory (LSTM) model, bidirectional long short-term memory (BiLSTM) model, and 1D convolutional neural network (CNN) model are used for sentiment analysis. The performance of each classifier is analyzed. The classification results showed that the features extracted from the tweets can be satisfactorily used to identify if a certain tweet is spam or not and create a learning model that will associate tweets with a particular sentiment.

Artificial intelligence has been applied in wildfire science and management since the 1990s, with early applications including neural networks and expert systems. Since then, the field has rapidly progressed congruently with the wide adoption of machine learning (ML) methods in the environmental sciences. Here, we present a scoping review of ML applications in wildfire science and management. Our overall objective is to improve awareness of ML methods among wildfire researchers and managers, as well as illustrate the diverse and challenging range of problems in wildfire science available to ML data scientists. To that end, we first present an overview of popular ML approaches used in wildfire science to date and then review the use of ML in wildfire science as broadly categorized into six problem domains, including (i) fuels characterization, fire detection, and mapping; (ii) fire weather and climate change; (iii) fire occurrence, susceptibility, and risk; (iv) fire behavior prediction; (v) fire effects; and (vi) fire management. Furthermore, we discuss the advantages and limitations of various ML approaches relating to data size, computational requirements, generalizability, and interpretability, as well as identify opportunities for future advances in the science and management of wildfires within a data science context. In total, to the end of 2019, we identified 300 relevant publications in which the most frequently used ML methods across problem domains included random forests, MaxEnt, artificial neural networks, decision trees, support vector machines, and genetic algorithms. As such, there exists opportunities to apply more current ML methods — including deep learning and agent-based learning — in the wildfire sciences, especially in instances involving very large multivariate datasets. We must recognize, however, that despite the ability of ML models to learn on their own, expertise in wildfire science is necessary to ensure realistic modelling of fire processes across multiple scales, while the complexity of some ML methods such as deep learning requires a dedicated and sophisticated knowledge of their application. Finally, we stress that the wildfire research and management communities play an active role in providing relevant, high-quality, and freely available wildfire data for use by practitioners of ML methods.

Decoding continuous 3D hand trajectories from electrocorticographic (ECoG) signals holds potential for brain-computer interface (BCI) applications. However, inter-session variability poses a major challenge for generalization. In this study, we propose a framework that leverages Riemannian-based feature extraction combined with stacked long short-term memory (LSTM) network to enable transfer learning across multiple sessions. ECoG recordings from five monkeys performing reaching tasks are considered. Spatial cross-frequency covariance matrices are computed over the brain area for each of 10 frequency band power and projected onto a Riemannian manifold to extract features which are invariant to session variability. These features and spectral feature are then used to train staked LSTM network. The results show that the proposed method achieves a stable cross-session performance and outperforms baseline models which are trained on frequency features. These findings highlight the potential of combining geometric features with temporal deep learning models for generalized decoding in translational BCI systems.

Smart cities require the development of information and communication technology to become a reality (ICT). A “smart city” is built on top of a “smart grid”. The implementation of numerous smart systems that are advantageous to the environment and improve the quality of life for the residents is one of the main goals of the new smart cities. In order to improve the reliability and sustainability of the transportation system, changes are being made to the way electric vehicles (EVs) are used. As EV use has increased, several problems have arisen, including the requirement to build a charging infrastructure, and forecast peak loads. Management must consider how challenging the situation is. There have been many original solutions to these problems. These heavily rely on automata models, machine learning, and the Internet of Things. Over time, there have been more EV drivers. Electric vehicle charging at a large scale negatively impacts the power grid. Transformers may face additional voltage fluctuations, power loss, and heat if already operating at full capacity. Without EV management, these challenges cannot be solved. A machine-learning (ML)-based charge management system considers conventional charging, rapid charging, and vehicle-to-grid (V2G) technologies while guiding electric cars (EVs) to charging stations. This operation reduces the expenses associated with charging, high voltages, load fluctuation, and power loss. The effectiveness of various machine learning (ML) approaches is evaluated and compared. These techniques include Deep Neural Networks (DNN), K-Nearest Neighbors (KNN), Long Short-Term Memory (LSTM), Random Forest (RF), Support Vector Machine (SVM), and Decision Tree (DT) (DNN). According to the results, LSTM might be used to give EV control in certain circumstances. The LSTM model’s peak voltage, power losses, and voltage stability may all be improved by compressing the load curve. In addition, we keep our billing costs to a minimum, as well.

Machine Learning (ML) technologies have been widely used for blind signal classification problem (Modulation Classification) to automatically identify the modulation schemes of Radio Frequency (RF) signal from complex- valued IQ (In-phase Quadrature) samples. Traditional ML methods usually have two stages where the first stage is to manually extract the features of the IQ symbols by subject matter experts and the second stage is to feed the features to an ML algorithm (e.g., a support vector machine) to develop the classifier. The state- of-art technology is to apply Deep Learning (DL) technologies such as Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN) directly to the complex-value IQ symbols to train a multi-class classifier. In this effort we are focused on the modulation and coding rate classification problems for multi-carrier Orthogonal Frequency-Division Multiplexing (OFDM) signals. The generated OFDM data set consists of over 480k instances across four different modulations and three different coding rates which include BPSK + 1=2, BPSK+3=4, QPSK+1=2, QPSK+3=4, QAM16+1=2, QAM16+3=4, QAM64+2=3 and QAM64+3=4. Deep Learning (DL) models can successfully catch the features of OFDM modulations and identify the modulation scheme from the baseband IQ symbols. Four DL models including Residual Neural Network 34(Resnet34), Resnet18, Squeezenet and Long Short-Term Memory (LSTM) trained over the data set can achieve over 94% overall accuracy for signals across 􀀀10dB to 10dB with step size of 4dB. Among them, Squeezenet and LSTM models have much smaller model sizes which can be easily loaded in resource-limited edge computing platforms. To the best of our knowledge, coding rate classification of OFDM signals has not been studied in previous works. Therefore, three different DL models including a single-stage coding rate classifier, a combination of modulation and coding rate classifier, and a two-stage coding rate classifier are developed to identify the coding rate. However, coding rates can only be identified for low-order modulation schemes BPSK and QPSK but not for high-order modulation schemes QAM16 and QAM64. Further investigation is needed to understand the coding rate classification for high-order QAM signals.

Application of machine learning in carbon capture and storage: An in-depth insight from the perspective of geoscience

Objective Negative emotions, such as stress and anger, are significant factors leading to dangerous driving behavior. Investigating the impact of these emotions on driving safety is crucial for effective traffic injury prevention. Methods Electroencephalography (EEG) is a valuable tool for detecting emotional neural activity due to its high temporal resolution and noninvasive characteristics. This study systematically analyzes the multiscale characteristics of EEG signals across different frequency bands, brain regions, and feature domains to elucidate the neural mechanisms of stress and anger states during driving. EEG signals were collected from 32 participants in a driving simulation environment under both emotional (stress and anger) and nonemotional conditions. First, time domain features (mean, variance, skewness, kurtosis), frequency domain features (power spectral density, Shannon entropy), and spatial domain features (common spatial patterns) were extracted from five key brain regions—frontal pole, frontal, central, parietal, and occipital—and across five frequency bands ranging from delta to gamma. Second, feature selection was performed through correlation analysis, and a support vector machine classifier was utilized for machine learning validation. Results The results indicate that the frontal pole region exhibited the highest frequency of neural activity, while frequency band analysis revealed that beta wave activity occurred most frequently. Furthermore, common spatial patterns demonstrated a significant ability to differentiate between stress and anger emotions. The optimal mixed feature set achieved classification accuracies of 85% for stress and 80% for anger, representing improvements of 20% to 55% and 7% to 55%, respectively, over the unselected feature set, as well as enhancements of 10% to 25% and 10% to 30%, respectively, compared to single-domain features. Conclusions These findings provide a neurophysiological basis for developing a wearable EEG-based driver emotion recognition system that has the potential to enhance traffic safety and reduce the risk of accidents by enabling timely interventions based on drivers’ emotional states.