Improve Model Performance Research Articles

Integrative factor-adjusted sparse generalized linear models

With the increasing accessibility of data, sample sizes remain limited in most studies, integrating multiple-source datasets is thus desirable to improve model performance. The high dependence among covariates is a common characteristic of high-dimensional data. However, traditional approaches fail to achieve variable selection consistency when the covariates are highly correlated. To address the issues stemming from correlation, we propose the penalized factor-adjusted approach to reduce the correlations in the integrative analysis within a generalized linear model (GLM) framework for multi-source high-dimensional data. By utilizing the latent factors and idiosyncratic components as predictors, the proposed approach enables model estimation with weakly correlated covariates. We rigorously establish the consistency properties under certain conditions. Simulation demonstrates the superior and competitive performance of the proposed approach in estimation, prediction, and variable selection. Analysis of genetic data on prostate cancer confirms its practical usefulness.

Prediction of Water-Richness Zoning of Weathered Bedrock Based on Whale Optimisation Algorithm and Random Forest

To effectively predict the water richness of weathered bedrock aquifers, the West First Plate area of the Hongliulin coal mine was taken as the study area, and 42 sets of pumping test borehole data from the weathered bedrock in the study area were used as training and testing samples. A total of five indicators related to the water richness of weathered bedrock, namely, the aquifer thickness, sand–base ratio, core take rate, degree of weathering, and lithological structure index, were selected. A prediction model for the water richness of weathered bedrock aquifers (WOA-RF) was subsequently proposed by combining the whale optimisation algorithm (WOA) and random forest (RF). This model can predict the water-richness level of weathered bedrock in an area with no pumping test data. The geological information from 98 sets of exploration boreholes in the study area was comprehensively used to achieve water-richness zoning of the weathered bedrock. The results indicated that the WOA is effective in optimising parameters and improving model performance. The accuracies of the optimal WOA-RF model in the training set and the test set were 93.1% and 92.3%, respectively. Compared with those of the single RF model, the accuracy, recall, and F1 value of the optimal WOA-RF model were increased by 11.3%, 18.2%, and 11%, respectively, and the differences before and after optimisation were obvious. A comparison and analysis of the predictive performance of each model revealed that the overall performance of the WOA-RF model was better than that of the other models. The weathered bedrock in the study area as a whole is weakly to moderately rich in water, and the predicted results are in good agreement with reality, which can provide a reference for future safe production in the West First Plate area.

Using machine learning and natural language processing in triage for prediction of clinical disposition in the emergency department

BackgroundAccurate triage is required for efficient allocation of resources and to decrease patients’ length of stay. Triage decisions are often subjective and vary by provider, leading to patients being over-triaged or under-triaged. This study developed machine learning models that incorporated natural language processing (NLP) to predict patient disposition. The models were assessed by comparing their performance with the judgements of emergency physicians (EPs).MethodThis retrospective study obtained data from patients visiting EDs between January 2018 and December 2019. Internal validation data came from China Medical University Hospital (CMUH), while external validation data were obtained from Asia University Hospital (AUH). Nontrauma patients aged ≥ 20 years were included. The models were trained using structured data and unstructured data (free-text notes) processed by NLP. The primary outcome was death in the ED or admission to the intensive care unit, and the secondary outcome was either admission to a general ward or transferal to another hospital. Six machine learning models (CatBoost, Light Gradient Boosting Machine, Logistic Regression, Random Forest, Extremely Randomized Trees, and Gradient Boosting) and one Logistic Regression derived from triage level were developed and evaluated using EPs’ predictions as reference.ResultA total of 17,2101 and 41,883 patients were enrolled from CMUH and AUH, respectively. EPs achieved F1 core of 0.361 and 0.498 for the primary and secondary outcomes, respectively. All machine learning models achieved higher F1 scores compared to EPs and Logistic Regression derived from triage level. Random Forest was selected for further evaluation and fine-tuning, because of its robust calibration and predictive performance. In internal validation, it achieved Brier scores of 0.072 and 0.089 for the primary and secondary outcomes, respectively, and 0.076 and 0.095 in external validation. Further analysis revealed that incorporating unstructured data significantly enhanced the model’s performance. Threshold adjustments were applied to improve clinical applicability, aiming to balance the trade-off between sensitivity and positive predictive value.ConclusionThis study developed and validated machine learning models that integrate structured and unstructured triage data to predict patient dispositions, distinguishing between general ward and critical conditions like ICU admissions and ED deaths. Integrating both structured and unstructured data significantly improved model performance.

Improved You Only Look Once v.8 Model Based on Deep Learning: Precision Detection and Recognition of Fresh Leaves from Yunnan Large-Leaf Tea Tree

Yunnan Province, China, known for its superior ecological environment and diverse climate conditions, is home to a rich resource of tea-plant varieties. However, the subtle differences in shape, color and size among the fresh leaves of different tea-plant varieties pose significant challenges for their identification and detection. This study proposes an improved YOLOv8 model based on a dataset of fresh leaves from five tea-plant varieties among Yunnan large-leaf tea trees. Dynamic Upsampling replaces the UpSample module in the original YOLOv8, reducing the data volume in the training process. The Efficient Pyramid Squeeze Attention Network is integrated into the backbone of the YOLOv8 network to boost the network’s capability to handle multi-scale spatial information. To improve model performance and reduce the number of redundant features within the network, a Spatial and Channel Reconstruction Convolution module is introduced. Lastly, Inner-SIoU is adopted to reduce network loss and accelerate the convergence of regression. Experimental results indicate that the improved YOLOv8 model achieves precision, recall and an mAP of 88.4%, 89.9% and 94.8%, representing improvements of 7.1%, 3.9% and 3.4% over the original model. This study’s proposed improved YOLOv8 model not only identifies fresh leaves from different tea-plant varieties but also achieves graded recognition, effectively addressing the issues of strong subjectivity in manual identification detection, the long training time of the traditional deep learning model and high hardware cost. It establishes a robust technical foundation for the intelligent and refined harvesting of tea in Yunnan’s tea gardens.

Spatio-temporal patterns and trends of streamflow in water-scarce Mediterranean basins

Abstract. The issue of water scarcity, exacerbated by climate change and demographic increase, has become a growing concern in many regions throughout the world. Understanding hydrological behaviour to promote resilient and sustainable water management is paramount. Hydrological models that integrate natural processes and anthropogenic alterations of the basin's hydrology are a powerful tool to support decision-making. We developed a SWAT+ hydrological model including stakeholder expert knowledge on water management and introducing a novel calibration and validation approach suitable for heterogeneous basins in space and/or time. We also assessed spatio-temporal patterns and trends of streamflow during the first 2 decades of the 21st century in the Catalan River Basin District, in the western Mediterranean, using a wide variety of indicators to fully characterize the hydrological regime. We calibrated and validated the model using data from 50 gauging stations, verifying the usefulness of the new calibration and validation strategy. Co-development with stakeholders and the integration of expert knowledge, most notably on reservoir operations, helped improve model performance. Results revealed a generalized streamflow reduction, as well as increased dominance of streamflow flashiness and zero-flow recurrence. We also observed differences in seasonal trends, with autumn being the most affected season. These results provide insights into how climate change and anthropogenic pressures are going to keep affecting water resources availability in the future, thus raising the need for sustainable management practices in the Catalan River Basin District, as well as other regions vulnerable to water scarcity.

Hybrid bagging and boosting with SHAP based feature selection for enhanced predictive modeling in intrusion detection systems

The novelty and growing sophistication of cyber threats mean that high accuracy and interpretable machine learning models are needed more than ever before for Intrusion Detection and Prevention Systems. This study aims to solve this challenge by applying Explainable AI techniques, including Shapley Additive explanations feature selection, to improve model performance, robustness, and transparency. The method systematically employs different classifiers and proposes a new hybrid method called Hybrid Bagging-Boosting and Boosting on Residuals. Then, performance is taken in four steps: the multistep evaluation of hybrid ensemble learning methods for binary classification and fine-tuning of performance; feature selection using Shapley Additive explanations values retraining the hybrid model for better performance and reducing overfitting; the generalization of the proposed model for multiclass classification; and the evaluation using standard information metrics such as accuracy, precision, recall, and F1-score. Key results indicate that the proposed methods outperform state-of-the-art algorithms, achieving a peak accuracy of 98.47% and an F1 score of 96.19%. These improvements stem from advanced feature selection and resampling techniques, enhancing model accuracy and balancing precision and recall. Integrating Shapley Additive explanations-based feature selection with hybrid ensemble methods significantly boosts the predictive and explanatory power of Intrusion Detection and Prevention Systems, addressing common pitfalls in traditional cybersecurity models. This study paves the way for further research on statistical innovations to enhance Intrusion Detection and Prevention Systems performance.

Mapping Soil Properties in Tropical Rainforest Regions Using Integrated UAV-Based Hyperspectral Images and LiDAR Points

For tropical rainforest regions with dense vegetation cover, the development of effective large-scale soil mapping methods is crucial to improve soil management practices to replace the time-consuming and laborious conventional approaches. While machine learning (ML) algorithms demonstrate superior predictability of soil properties over linear models, their practical and automated application for predicting soil properties using remote sensing data requires further assessment. Therefore, this study aims to integrate Unmanned Aerial Vehicles (UAVs)-based hyperspectral images and Light Detection and Ranging (LiDAR) points to predict the soil properties indirectly in two tropical rainforest mountains (Diaoluo and Limu) in Hainan Province, China. A total of 175 features, including texture features, vegetation indices, and forest parameters, were extracted from two study sites. Six ML models, Partial Least Squares Regression (PLSR), Random Forest (RF), Adaptive Boosting (AdaBoost), Gradient Boosting Decision Trees (GBDT), Extreme Gradient Boosting (XGBoost), and Multilayer Perceptron (MLP), were constructed to predict soil properties, including soil acidity (pH), total nitrogen (TN), soil organic carbon (SOC), and total phosphorus (TP). To enhance model performance, a Bayesian optimization algorithm (BOA) was introduced to obtain optimal model hyperparameters. The results showed that compared with the default parameter tuning method, BOA always improved models’ performances in predicting soil properties, achieving average R2 improvements of 202.93%, 121.48%, 8.90%, and 38.41% for soil pH, SOC, TN, and TP, respectively. In general, BOA effectively determined the complex interactions between hyperparameters and prediction features, leading to an improved model performance of ML methods compared to default parameter tuning models. The GBDT model generally outperformed other ML methods in predicting the soil pH and TN, while the XGBoost model achieved the highest prediction accuracy for SOC and TP. The fusion of hyperspectral images and LiDAR data resulted in better prediction of soil properties compared to using each single data source. The models utilizing the integration of features derived from hyperspectral images and LiDAR data outperformed those relying on one single data source. In summary, this study highlights the promising combination of UAV-based hyperspectral images with LiDAR data points to advance digital soil property mapping in forested areas, achieving large-scale soil management and monitoring.

Interaction of Asymmetric Adaptive Network Structures and Parameter Balance in Image Feature Extraction and Recognition

To better process irregular sample images for their image feature extraction and recognition, this essay proposes asymmetric adaptive neural network (AACNN) structures, including dual structures of an adaptive image feature extraction network (AT-CNN) and adaptive image recognition network (AT-ACNN). They both comprise an Adaptive Transform (AT) module and a deep learning network, but the ACNN comprises pixel-adaptive convolutional (PAC) kernels that CNN does not have, reflecting the asymmetry of these network structures. Structural analysis and comparative testing experiments indicated that the proposed method is more appropriate and effective for dealing with irregular sample images with different sizes and views, mainly focusing on their feature extraction accuracy and image recognition efficiency. The proposed method constructs the interaction between asymmetric dual network structures, essential in improving model performance and efficiency. It specifically manifests that the PAC kernels in an ACNN resolves the problem of content-agnostic convolution in image recognition by learning image features from a pre-trained CNN. On the other hand, it improves image recognition efficiency by using feature maps extracted from the pre-trained CNN to train the classifiers in the ACNN. We also found that parameter balance was essential in adaptive neural network structure for better performance in further testing experiments. When setting the Dropout layer parameter at 0.5 and the iteration number was 32, the proposed model achieved adequate recognition accuracy and efficiency. Smaller parameters affect model performance, but more extensive parameters significantly increase computational burden and loss. Comparative testing experiments fully validated its superiority compared with traditional methods based on CNNs. Using traditional carving patterns from Anhui Province as an example, we conducted image recognition and feature graphic application under ideal parameter balance conditions and thereby demonstrated the practicality and value of the proposed method.

A text classification method combining in-domain pre-training and prompt learning for the steel e-commerce industry

Purpose With the development of Web information systems, steel e-commerce platforms have accumulated a large number of quality objection texts. These texts reflect consumer dissatisfaction with the dimensions, appearance and performance of steel products, providing valuable insights for product improvement and consumer decision-making. Currently, mainstream solutions rely on pre-trained models, but their performance on domain-specific data sets and few-shot data sets is not satisfactory. This paper aims to address these challenges by proposing more effective methods for improving model performance on these specialized data sets. Design/methodology/approach This paper presents a method on the basis of in-domain pre-training, bidirectional encoder representation from Transformers (BERT) and prompt learning. Specifically, a domain-specific unsupervised data set is introduced into the BERT model for in-domain pre-training, enabling the model to better understand specific language patterns in the steel e-commerce industry, enhancing the model’s generalization capability; the incorporation of prompt learning into the BERT model enhances attention to sentence context, improving classification performance on few-shot data sets. Findings Through experimental evaluation, this method demonstrates superior performance on the quality objection data set, achieving a Macro-F1 score of 93.32%. Additionally, ablation experiments further validate the significant advantages of in-domain pre-training and prompt learning in enhancing model performance. Originality/value This study clearly demonstrates the value of the new method in improving the classification of quality objection texts for steel products. The findings of this study offer practical insights for product improvement in the steel industry and provide new directions for future research on few-shot learning and domain-specific models, with potential applications in other fields.

Predicting future fallers in Parkinson’s disease using kinematic data over a period of 5 years

Parkinson’s disease (PD) increases fall risk, leading to injuries and reduced quality of life. Accurate fall risk assessment is crucial for effective care planning. Traditional assessments are subjective and time-consuming, while recent assessment methods based on wearable sensors have been limited to 1-year follow-ups. This study investigated whether a short sensor-based assessment could predict falls over up to 5 years. Data from 104 people with PD without prior falls were collected using six wearable sensors during a 2-min walk and a 30-s postural sway task. Five machine learning classifiers analysed the data. The Random Forest classifier performed best, achieving 78% accuracy (AUC = 0.85) at 60 months. Most models showed excellent performance at 24 months (AUC > 0.90, accuracy 84–92%). Walking and postural variability measures were key predictors. Adding clinicodemographic data, particularly age, improved model performance. Wearable sensors combined with machine learning can effectively predict fall risk, enhancing PD management and prevention strategies.

An Evaluation of Image Slicing and YOLO Architectures for Object Detection in UAV Images

Object detection in aerial images poses significant challenges due to the high dimensions of the images, requiring efficient handling and resizing to fit object detection models. The image-slicing approach for object detection in aerial images can increase detection accuracy by eliminating pixel loss in high-resolution image data. However, determining the proper dimensions to slice is essential for the integrity of the objects and their learning by the model. This study presents an evaluation of the image-slicing approach for alternative sizes of images to optimize efficiency. For this purpose, a dataset of high-resolution images collected with Unmanned Aerial Vehicles (UAV) has been used. The experiments evaluated using alternative YOLO architectures like YOLOv7, YOLOv8, and YOLOv9 show that the image dimensions significantly change the performance results. According to the experiments, the best mAP@05 accuracy was obtained by slicing 1280×1280 for YOLOv7 producing 88.2. Results show that edge-related objects are better preserved as the overlap and slicing sizes increase, resulting in improved model performance.

Feature selection and response prediction on a suspension bridge due to wind effect by machine learning

This research introduces an efficient machine learning approach for predicting the structural health of long-span suspension bridges, focusing on their responses to dynamic environmental load like earthquakes and wind. Unlike traditional analytical methods that rely on simplifications, this study utilizes real-world data from anemometers and accelerometers to enhance prediction accuracy and efficiency. Also instead of using whole signals, the study uses fewer time frame signal. The main contribution of this paper is the application of support vector regression (SVR), optimized through Observer-Teacher-Learner-Based-Optimization (OTLBO) for parameter tuning. The optimization process is further refined using Multi-Objective Observer-Teacher-Learner-Based-Optimization (MOOTLBO), targeting feature selection and dimension reduction to improve model performance significantly. Hardanger Bridge serves as a case study, demonstrating the model's capability to accurately predict acceleration responses to wind. By inputting wind sensor data and analyzing acceleration outputs, the enhanced SVR model, through OTLBO and MOOTLBO, shows remarkable predictive accuracy, validated against six performance indices. This methodological advancement suggests that the machine learning-based model could potentially supplant traditional finite element models, offering a more adaptable, efficient, and accurate tool for structural health monitoring (SHM). This advancement addresses critical discrepancies in SHM systems, such as data gaps or sensor malfunctions, by providing a reliable prediction method that ensures the ongoing safety and integrity of bridge structures.

Study on bionics-based swarm intelligence optimization algorithms for wavelength selection in near-infrared spectroscopy

Wavelength selection is one of the most important steps in the modeling of near-infrared spectroscopy (NIRS), which is of great significance to reduce model complexity and improve model performance. In this paper, a total of ten bionics-based swarm intelligence optimization algorithms (BSIOAs) inspired by natural creatures, such as Harris Hawks Optimization (HHO), Butterfly Optimization Algorithm (BOA), Whale Optimization Algorithm (WOA), Monarch Butterfly Optimization (MBO), Grey Wolf Optimization (GWO), Fruit Fly Optimization Algorithm (FOA), Bat Algorithm (BA), Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Genetic Algorithm (GA) were studied on application to wavelength selection in the NIRS modeling. Three benchmark NIRS datasets were used to evaluate the algorithms by calculating the indicators, including coefficients of determination, root mean square error, and residual predictive deviation in calibration and prediction. The results obtained showed that these BSIOAs can significantly reduce the number of wavelengths (retaining half or fewer). Compared with the full-spectrum models, the present models not only simplified the model structures but improved the model performances. The performances were generally better than the ones by some popular and classic wavelength selection algorithms, such as competitive adaptive reweighted sampling, Monte Carlo uninformative variable elimination, variable importance in projection, interval partial least-squares, and successive projections algorithm.

Intrusion Detection in IoT using Gaussian Fuzzy Mutual Information-based Feature Selection

The proliferation of Internet of Things (IoT) devices has revolutionized various sectors by enabling real-time monitoring, data collection, and intelligent decision-making. However, the massive volume of data generated by these devices presents significant challenges for data processing and analysis. Intrusion Detection Systems (IDS) for IoT require efficient and accurate identification of malicious activities amidst vast amounts of data. Feature selection is a critical step in this process, aiming to identify the most relevant features that contribute to accurate intrusion detection, thus reducing computational complexity and improving model performance. Traditional Mutual Information-based Feature Selection (MIFS) methods face challenges when applied to IoT data due to their inherent noise, uncertainty, and imprecision. This study introduces a novel Fuzzy Mutual Information-based Feature Selection (Fuzzy-MIFS) method that integrates fuzzy logic with Gaussian membership functions to address these challenges. The proposed method enhances the robustness and effectiveness of the feature selection process, resulting in improved accuracy and efficiency of IDSs in IoT environments. Experimental results demonstrate that the Fuzzy-MIFS method consistently outperformed existing feature selection techniques across various neural network models, such as CNN, LSTM, and DBN, showcasing its superior performance in handling the complexities of IoT data. The results show that Fuzzy-MIFS increased the accuracy from 0.962 to 0.986 for CNN, from 0.96 to 0.968 for LSTM, and from 0.96 to 0.97 for DBN.

Diel temperature patterns unveiled: High-frequency monitoring and deep learning in lake Kasumigaura

The diel variation of water temperatures is crucial information for lakes. This study investigated such variation in Lake Kasumigaura, the second largest lake in Japan, situated 60 km from the Tokyo metropolitan area, utilizing high-frequency monitoring and deep learning techniques. Data from high-frequency monitoring of vertical water temperatures at seven depths were employed. A convolutional neural network (CNN) based deep learning model was developed and assessed across three input scenarios. The impact of forecast horizons ranging from one to 48 h ahead on model performance was examined. Results indicate a degradation in model performance with increasing forecast horizons, irrespective of input scenario or water depth. Notably, the CNN model demonstrates superior performance in near-term and medium-term forecasts compared to long-term predictions, underscoring the need for enhanced efforts in long-term forecasting. Overall, the CNN model effectively reproduces vertical water temperature profiles and captures diel variations in lake water temperatures. Incorporating additional input variables does not necessarily improve model performance; however, using surface water temperatures and air temperatures as inputs produces acceptable results for modeling vertical temperature profiles. These findings have implications for lake management in Lake Kasumigaura (e.g., controlling phytoplankton and zooplankton communities) and offer insights into water temperature modeling for other lakes.

The predictivity of QSARs for toxicity: Recommendations for improving model performance

The predictivity of QSARs for toxicity: Recommendations for improving model performance

A probabilistic approach driven credit card anomaly detection with CBLOF and isolation forest models

Businesses throughout the world are starting to accept more and more electronic payment alternatives for in-person and online purchases. Unusual behaviour such as internet fraud and payment defaults, which may lead to significant monetary losses, have become increasingly widespread as credit card use in online buying has expanded. To find a solution to this problem, researchers looked into a number of different machine learning classifiers that may identify irregularities in the data pertaining to credit card transactions. Overlapping class samples and an uneven distribution of classes make it hard to find outliers in this data. General learning algorithms may favour samples from the majority class, and the detection rate of anomalies in samples from minority classes is thus low. Our proposed Credit Card Outlier Detection (CCOD) model incorporates many machine learning algorithms to enhance the detection rates of credit card anomalies. Utilising the stratified sampling technique and the k-fold cross-validation process, we address data imbalance and overfitting. Working with optimized selected features rather than all features reduces the danger of overfitting and improves model performance. It has been noted that the findings of Cluster-Based Local Outlier Factor (CBLOF) classifier on European Credit Card Dataset and Isolation Forest classifier on German credit card dataset were superior to those of the other classifiers with Mathew correlation coefficient value of 0.95 and 0.97 respectively. The CBLOF, an advanced outlier detection method that evaluates the local density of data points within clusters. By comparing the density of a point to the densities of its surrounding clusters, CBLOF identifies outliers based on their deviation from normal cluster behaviour. In general, this CCOD model shows that it is better at dealing with issues like uneven class distribution, overfitting, and classes that overlap.

Detection and Classification of Malicious Websites Using Natural Language Processing (NLP) and Machine Learning (ML) Techniques

The internet, while offering extensive services and information, has also become a platform for malicious activities, particularly through harmful websites that threaten cybersecurity. Detecting and classifying these websites is crucial for protecting users from online threats. Traditional detection methods, primarily based on blacklists and signature-based techniques, struggle to match the pace with the dynamic evolving strategies of cybercriminals. Recent advancements in Machine Learning (ML) show promise, though they remain works in progress. This research addressed this challenge by exploring the usage of Natural Language Processing and Machine Learning techniques used to classify websites as benign or malicious. Unlike many existing studies that relied on URL features alone, this study incorporated a more comprehensive feature set, including URL, content, and additional web attributes, which enhanced classification accuracy. Using an imbalanced dataset skewed towards malicious sites, this study solved using SMOTE (Synthetic Minority Over-sampling Technique) the class imbalance problem, improving model performance. Utilized Hashing Vectorizer (HashingV) and TF-IDF (Term Frequency-Inverse Document Frequency), were adopted to transform textual features into their vector representations while PCA (Principal Component Analysis) and truncated Singular Value Decomposition (truncSVD), were then used to optimize feature representation across different dimensions. Five ML classifiers include Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), K-Nearest Neighbors (KNN), and Logistic Regression (LR) were tested for classification, and performance was evaluated using metrics such as precision and recall, accuracy, F1-Score. The results revealed that Random Forest classifier utilizing HashingV recorded the best results, with accuracies of 99.9563% using truncSVD and 99.9561% with PCA.

Multichannel EMG-based gesture recognition utilizing advanced machine learning techniques: A random forest classifier for high-precision signal classification

This research examines how advanced machine learning algorithms can be used to classify multichannel electromyographic (EMG) signals with a high level of accuracy to assist in recognizing hand gestures. The goal is to create a robust and scalable system for gesture-based virtual control using EMG signals with potential applications in assistive technologies, rehabilitation, and human-computer interaction. Data were gathered using a MYO Thalmic bracelet containing eight EMG sensors on thirty-six subjects, and a Random Forest classifier was trained to identify seven distinct types of hand gestures (rest, fist clench, wrist flexion/extension, and radial/ulnar deviations). The machine learning pipeline included extensive preprocessing (i.e., EMG signal normalization and signal feature extraction; root mean square, waveform length, and zero crossing rate) and several hyperparameter tuning procedures to improve model performance. The Random Forest model (100 decision trees) achieved an overall classification accuracy of 98.68%, with a range of accuracies for each class (e.g., 95.2% wrist flexion and 91.8% ulnar deviation) when evaluated using cross-validation (i.e., average F1-score = 0.92, precision = 0.94, recall = .91). Overall, the study provides strong evidence for the effectiveness of ensemble learning methods at analyzing complex, multidimensional EMG signals. The high classification accuracy reported, in particular, demonstrates that the system could function for real-time recognition of hand gestures in a virtual environment. Ultimately, the initial work sets the stage for future exploration of a model that may be integrated with actuation models to control prosthetic limbs, virtual actors/avatars, and robotic devices. By demonstrating a scalable and efficient method of gesture recognition using EMG signals, these early findings enable future pathways and possibilities to design innovative, assistive solutions for digital systems that increase accessibility and interaction for users who are motor impaired or have a limited range of motion.

Exploration of Imbalanced Regression in state-of-health estimation of Lithium-ion batteries

The state of health (SOH) estimation for lithium-ion batteries based on deep learning (DL) has made great progress. However, due to different electrochemical compositions of lithium-ion batteries, different ways of conducting experiments and other factors, the degradation process of some batteries shows longer early degradation time and shorter later degradation time, resulting in a long-tailed distribution of degradation data. This leads to the problem of data imbalance in SOH estimation tasks, which affects the accuracy of SOH estimation. This article explores the long-tailed distribution phenomenon in the field of batteries and the corresponding imbalanced regression problem it brings to the estimation of battery SOH. In addition, a method for improving model performance is proposed. Specifically, we use a quadratic interpolation and standardization method to analyze the battery data to ensure the consistency of data features. By discretized analysis of continuous problems, the label distribution smoothing (LDS) method is applied to deep neural networks to analyze and solve this imbalanced regression problem. By convolution processing with the kernel function and label distribution, the weights corresponding to different labels are calculated, which improves the estimation accuracy. We conducted battery aging experiments and verified that the degradation data follows a long-tailed distribution. The effectiveness of the final method was validated on our experimental data and a publicly available dataset.