Search

Micro, Small, and Medium-sized Enterprises (MSMEs) are essential for the growth and development of the country's economy, as they create jobs, generate income, and foster production and innovation. In recent years, credit risk assessment (CRA) has been an essential process used by financial institutions to evaluate the creditworthiness of MSMEs and determine the likelihood of default. Traditionally, CRA has relied on credit scores and financial statements, but with the advent of machine learning (ML) algorithms, lenders have a new tool at their disposal. By and large, ML algorithms are designed to classify borrowers based on their credit history and transactional data while leveraging the entity relationship involved in credit transactions. This study introduces an innovative knowledge graph-driven credit risk assessment model (RGCN-RF) based on the Relational Graph Convolutional Network (RGCN) and Random Forest (RF) algorithm. RGCN is employed to identify topological structures and relationships, which is currently nascent in traditional credit risk assessment methods. RF categorises MSMEs based on the enterprise embedding vector generated from RGCN. Extensive experimentation is conducted to assess model performance utilising the Indian MSMEs database. The balanced accuracy of 92% obtained using the RGCN-RF model demonstrates a considerable advancement over prior techniques in identifying risk-free enterprises. [ABSTRACT FROM AUTHOR]

Two-dimensional MXenes materials are expected to be cost-effective and efficient catalysts for hydrogen evolution reaction (HER) due to their tunable surface electronic structure. To optimize its catalytic activity, this study proposes a data-driven framework (CROSST) to generate descriptors through feature engineering and integrate them into a convolutional neural network to predict the HER performance of transition-metal (TM) anchored bis-transition-metal carbon-nitride (TM-M′ 2 M′′CNO 2) catalysts. The STCNN model achieves a coefficient of determination of more than 0.93, showing the ability to predict the Gibbs free energy of hydrogen adsorption (Δ G H) with high accuracy. We explored the microscopic mechanisms of HER activity through density functional theory and machine learning analysis. It was found that TM anchoring altered the charge distribution of the MXenes material, elevated the atomic orbital occupancy of neighboring O sites, and weakened the O adsorption of H, thus affecting the HER activity. These results provide a theoretical basis for the design of high-performance MXenes HER catalysts and contribute new research perspectives. [Display omitted] • Propose a CNN framework using random forest and symbol transformer for optimization. • Examine the impact of transition metal anchoring on MXenes catalytic performance. • MXenes catalytic behavior is controlled by the outermost and anchoring metals. [ABSTRACT FROM AUTHOR]

Credit card fraud is a huge challenge in the financial sector, causing huge losses every year. The problem is exacerbated by increased marketing and sophisticated fraudulent activities. This study addresses the important issue of accurate real-time detection of fraudulent transactions to minimize financial losses and enhance transactional security. The main objective of this study is to develop a comprehensive fraud detection algorithm using deep learning techniques, specially designed to address the complexity and volume of modern credit card transactions. Key contributions of this research include the presentation of a new deep learning algorithm optimized for credit card fraud detection, the integration of feature engineering techniques to improve the performance of the model, and a potential scalable solution analysis in real-time Significant improvement in proven rates. The results show that the proposed deep learning-based model achieves higher accuracy and lower false positive rate, giving financial institutions a significant advantage in protecting against fraudulent activities about the character. This study highlights the power of deep learning in reforming fraud detection systems, and lays the foundation for future developments in this important area. [ABSTRACT FROM AUTHOR]

Breast cancer (BC) is a prominent cause of female mortality on a global scale. Recently, there has been growing interest in utilizing blood and tissue-based biomarkers to detect and diagnose BC, as this method offers a non-invasive approach. To improve the classification and prediction of BC using large biomarker datasets, several machine-learning techniques have been proposed. In this paper, we present a multi-stage approach that consists of computing new features and then sorting them into an input image for the ResNet50 neural network. The method involves transforming the original values into normalized values based on their membership in the Gaussian distribution of healthy and BC samples of each feature. To test the effectiveness of our proposed approach, we employed the Coimbra and Wisconsin datasets. The results demonstrate efficient performance improvement, with an accuracy of 100% and 100% using the Coimbra and Wisconsin datasets, respectively. Furthermore, the comparison with existing literature validates the reliability and effectiveness of our methodology, where the normalized value can reduce the misclassified samples of ML techniques because of its generality. [ABSTRACT FROM AUTHOR]

Amid growing interest in the precise detection of volatile organic compounds (VOCs) in industrial field, the demand for highly effective gas sensors is at an all‐time high. However, traditional sensors with their classic single‐output signal, bulky and complex integrated structure when forming array often involve complicated technology and high cost, limiting their widespread adoption. Here, this study introduces a novel approach, employing an integrated YSZ‐based (YSZ: yttria‐stabilized zirconia) mixed potential sensor equipped with a triple‐sensing electrode array, to efficiently detect and differentiate six types of VOCs gases. This innovative sensor integrates NiSb2O6, CuSb2O6, and MgSb2O6 sensing electrodes (SEs), which are sensitive to pentane, isoprene, n‐propanol, acetone, acetic acid, and formaldehyde gases. Through feature engineering based on intuitive spike‐based response values, it accentuates the distinct characteristics of every gas. Eventually, an average classification accuracy of 98.8% and an overall R‐squared error (R2) of 99.3% for concentration regression toward six target gases can be achieved, showcasing the potential to quantitatively distinguish between industrial hazardous VOCs gases. [ABSTRACT FROM AUTHOR]

The ever-growing 1.15 million new cases of cancer on a yearly basis alone in India is a major cause of concern for the experts and researchers working in various biomedical organizations. The advent of modern text engineering strategies and NLP techniques can play a crucial role in the discovery and analysis of pre-existing knowledge present in the cancer related biomedical archives. The available 10 Cancer hallmarks can provide key insights and make significant impact in the ongoing cancer research. It is extremely important to identify and classify required information due to time and resource crunch which needs to be quickly accessed. This article introduces a novel machine learning framework called Cancer Hallmark Classification and Topic Modeling (CHCTM), designed for supervised learning. The CHCTM framework is capable of semantically learning, categorizing, and extracting significant topics and their combinations related to the hallmarks of cancer (HoC) from a dataset comprising 1499 PubMed documents. The key contributions of this research include the creation of an innovative ensemble classification model using a meta-classifier based on Random Forest (RF). Additionally, it introduces an Enhanced Latent Dirichlet Allocation (ELDA) topic modeling strategy to generate relevant mixtures of topics. The performance of the CHCTM framework is evaluated using precision, recall, accuracy, and F-score parameters. Comparative analysis with other biomedical baseline methods reveals an 8% improvement in F-score. The coherence values acquired for ELDA are tallied and weighted against PLSA and LDA models to demonstrate the effectiveness of this approach. [ABSTRACT FROM AUTHOR]

The agricultural industry is critical to the Indian GDP, yet it faces several issues, such as energy, unemployment, and water scarcity. Among these challenges, reliable access to electricity stands out as a crucial factor for implementing smart farming techniques tailored to diverse crops and irrigation patterns. This paper addresses the intricate task of load forecasting in the agriculture sector, employing machine learning and ensemble learning methods. By scrutinizing various agroclimatic and weather characteristics, we pinpoint exogenous features influencing load prediction and prioritize their significance. The study introduces a groundbreaking approach, Temporal Adaptive Ensemble Learning (TAEL), designed to augment accuracy by incorporating temporal dynamics. Six distinct cases, varying in input parameters, are meticulously examined, employing ensemble approaches and benchmarking against Multiple Linear Regression (MLR) and K-Nearest Neighbor (KNN) models. The findings emphasize the vital role of integrating agroclimatic and time series features, showcasing notable enhancements in algorithmic performance across all cases. The proposed TAEL model stands out, achieving a remarkable mean accuracy of 87.1% and a 20.4% MAPE error. This underscores its effectiveness compared to traditional statistical and machine learning approaches, addressing limitations observed in existing studies. The research not only identifies critical factors influencing load prediction but also pioneers a novel approach, TAEL, contributing to the advancement of agricultural load forecasting methodologies. [ABSTRACT FROM AUTHOR]

Flotation is a crucial separation technique in coal beneficiation. However, the inability to quickly and accurately detect ash content during the flotation process hinders the development of flotation automation. This paper introduced a novel hybrid model based on convolutional neural networks, establishing the ash content in tailings. In this method, convolutional neural network is utilized to automatically extract features from tailings images. In the model training and optimization phase after inputting the augmented data, the optimal ResNet18 model is identified by comparing five model optimization methods and replacing the SoftMax classifier with the traditional SVR in ResNet18. Additionally, a PCA-SVR model for predicting tailings ash content is constructed for comparative analysis, employing six different feature combinations to validate the effectiveness of the proposed features, with optimal support vector parameters identified through random search and cross-validation algorithms. The prediction results show that the PCA-SVR model has the highest accuracy when the feature dimension is reduced to 16, and the prediction of ResNet18-SVR is more accurate compared to PCA-SVR. This indicates that the proposed method can predict the ash content more accurately than the traditional feature engineering methods and has a broad application prospect in the field of coal beneficiation. [ABSTRACT FROM AUTHOR]

Thermal Energy Storage (TES) plays a pivotal role in the fire protection of Li-ion batteries, especially for the high-voltage (HV) battery systems in Electrical Vehicles (EVs). This study covers the application of TES in mitigating thermal runaway risks during different battery charging/discharging conditions known as Vehicle-to-grid (V2G) and Grid-to-vehicle (G2V). Through controlled simulations in Simulink, this research models real-world scenarios to analyze the effectiveness of TES in controlling battery conditions under various environmental conditions. This study also integrates Machine Learning (ML) techniques to utilize the produced data by the simulation model and to predict any probable thermal spikes and enhance the system reliability, focusing on crucial factors like battery temperature, current, or State of charge (SoC). Feature engineering is also employed to identify the key parameters among all features that are considered for this study. For a broad comparison among different models, three different ML techniques, logistic regression, support vector machine (SVM), and Naïve Bayes, have been used alongside their hybrid combination to determine the most accurate one for the related topic. This study concludes that SoC is the most significant factor affecting thermal management while grid power consumption has the least impact. Additionally, the findings demonstrate that logistic regression outperforms other methods, with the improving feature to be used in the hybrid models as it can increase their efficiency due to its linearity capture capability. [ABSTRACT FROM AUTHOR]

Prevalence estimates of Parkinson's disease (PD)—the fastest-growing neurodegenerative disease—are generally underestimated due to issues surrounding diagnostic accuracy, symptomatic undiagnosed cases, suboptimal prodromal monitoring, and limited screening access. Remotely monitored wearable devices and sensors provide precise, objective, and frequent measures of motor and non-motor symptoms. Here, we used consumer-grade wearable device and sensor data from the WATCH-PD study to develop a PD screening tool aimed at eliminating the gap between patient symptoms and diagnosis. Early-stage PD patients (n = 82) and age-matched comparison participants (n = 50) completed a multidomain assessment battery during a one-year longitudinal multicenter study. Using disease- and behavior-relevant feature engineering and multivariate machine learning modeling of early-stage PD status, we developed a highly accurate (92.3%), sensitive (90.0%), and specific (100%) random forest classification model (AUC = 0.92) that performed well across environmental and platform contexts. These findings provide robust support for further exploration of consumer-grade wearable devices and sensors for global population-wide PD screening and surveillance. [ABSTRACT FROM AUTHOR]

High-quality short-term forecasts of electrical energy generation in solar power plants are crucial in the dynamically developing sector of renewable power generation. This article addresses the issue of selecting appropriate (preferred) methods for forecasting energy generation from a solar power plant within a 15 min time horizon. The effectiveness of various machine learning methods was verified. Additionally, the effectiveness of proprietary ensemble and hybrid methods was proposed and examined. The research also aimed to determine the appropriate sets of input variables for the predictive models. To enhance the performance of the predictive models, proprietary additional input variables (feature engineering) were constructed. The significance of individual input variables was examined depending on the predictive model used. This article concludes with findings and recommendations regarding the preferred predictive methods. [ABSTRACT FROM AUTHOR]

Electroencephalogram (EEG) signals contain information about the brain's state as they reflect the brain's functioning. However, the manual interpretation of EEG signals is tedious and time-consuming. Therefore, automatic EEG translation models need to be proposed using machine learning methods. In this study, we proposed an innovative method to achieve high classification performance with explainable results. We introduce channel-based transformation, a channel pattern (ChannelPat), the t algorithm, and Lobish (a symbolic language). By using channel-based transformation, EEG signals were encoded using the index of the channels. The proposed ChannelPat feature extractor encoded the transition between two channels and served as a histogram-based feature extractor. An iterative neighborhood component analysis (INCA) feature selector was employed to select the most informative features, and the selected features were fed into a new ensemble k-nearest neighbor (tkNN) classifier. To evaluate the classification capability of the proposed channel-based EEG language detection model, a new EEG language dataset comprising Arabic and Turkish was collected. Additionally, Lobish was introduced to obtain explainable outcomes from the proposed EEG language detection model. The proposed channel-based feature engineering model was applied to the collected EEG language dataset, achieving a classification accuracy of 98.59%. Lobish extracted meaningful information from the cortex of the brain for language detection. [ABSTRACT FROM AUTHOR]

Neurodegenerative diseases often exhibit a strong link with sleep disruption, highlighting the importance of effective sleep stage monitoring. In this light, automatic sleep stage classification (ASSC) plays a pivotal role, now more streamlined than ever due to the advancements in deep learning (DL). However, the opaque nature of DL models can be a barrier in their clinical adoption, due to trust concerns among medical practitioners. To bridge this gap, we introduce SleepBoost, a transparent multi-level tree-based ensemble model specifically designed for ASSC. Our approach includes a crafted feature engineering block (FEB) that extracts 41 time and frequency domain features, out of which 23 are selected based on their high mutual information score (> 0.23). Uniquely, SleepBoost integrates three fundamental linear models into a cohesive multi-level tree structure, further enhanced by a novel reward-based adaptive weight allocation mechanism. Tested on the Sleep-EDF-20 dataset, SleepBoost demonstrates superior performance with an accuracy of 86.3%, F1-score of 80.9%, and Cohen kappa score of 0.807, outperforming leading DL models in ASSC. An ablation study underscores the critical role of our selective feature extraction in enhancing model accuracy and interpretability, crucial for clinical settings. This innovative approach not only offers a more transparent alternative to traditional DL models but also extends potential implications for monitoring and understanding sleep patterns in the context of neurodegenerative disorders. The open-source availability of SleepBoost's implementation at https://github.com/akibzaman/SleepBoost can further facilitate its accessibility and potential for widespread clinical adoption. [ABSTRACT FROM AUTHOR]

The article describes a new method for malware classification, based on a Machine Learning (ML) model architecture specifically designed for malware detection, enabling real-time and accurate malware identification. Using an innovative feature dimensionality reduction technique called the Interpolation-based Feature Dimensionality Reduction Technique (IFDRT), the authors have significantly reduced the feature space while retaining critical information necessary for malware classification. This technique optimizes the model's performance and reduces computational requirements. The proposed method is demonstrated by applying it to the BODMAS malware dataset, which contains 57,293 malware samples and 77,142 benign samples, each with a 2381-feature vector. Through the IFDRT method, the dataset is transformed, reducing the number of features while maintaining essential data for accurate classification. The evaluation results show outstanding performance, with an F1 score of 0.984 and a high accuracy of 98.5% using only two reduced features. This demonstrates the method's ability to classify malware samples accurately while minimizing processing time. The method allows for improving computational efficiency by reducing the feature space, which decreases the memory and time requirements for training and prediction. The new method's effectiveness is confirmed by the calculations, which indicate significant improvements in malware classification accuracy and efficiency. The research results enhance existing malware detection techniques and can be applied in various cybersecurity applications, including real-time malware detection on resource-constrained devices. Novelty and scientific contribution lie in the development of the IFDRT method, which provides a robust and efficient solution for feature reduction in ML-based malware classification, paving the way for more effective and scalable cybersecurity measures. [ABSTRACT FROM AUTHOR]

Understanding the correlation between the fundamental descriptors and catalytic performance is meaningful to guide the design of high‐performance electrochemical catalysts. However, exploring key factors that affect catalytic performance in the vast catalyst space remains challenging for people. Herein, to accurately identify the factors that affect the performance of N2 reduction, we apply interpretable machine learning (ML) to analyze high‐throughput screening results, which is also suited to other surface reactions in catalysis. To expound on the paradigm, 33 promising catalysts are screened from 168 carbon‐supported candidates, specifically single‐atom catalysts (SACs) supported by a BC3 monolayer (TM@VB/C‐Nn = 0–3‐BC3) via high‐throughput screening. Subsequently, the hybrid sampling method and XGBoost model are selected to classify eligible and non‐eligible catalysts. Through feature interpretation using Shapley Additive Explanations (SHAP) analysis, two crucial features, that is, the number of valence electrons (Nv) and nitrogen substitution (Nn), are screened out. Combining SHAP analysis and electronic structure calculations, the synergistic effect between an active center with low valence electron numbers and reasonable C‐N coordination (a medium fraction of nitrogen substitution) can exhibit high catalytic performance. Finally, six superior catalysts with a limiting potential lower than −0.4 V are predicted. Our workflow offers a rational approach to obtaining key information on catalytic performance from high‐throughput screening results to design efficient catalysts that can be applied to other materials and reactions. [ABSTRACT FROM AUTHOR]

This article discusses the creation of a prototype for a turbine digital twin, specifically focusing on developing a solution for constructing and continuously updating a mathematical model of the stationary operating modes of a heating turbine. A significant challenge in creating turbine digital twins is that the equipment often operates for extended periods within a limited range of permissible parameters. However, the model must be updated to cover the entire range of operating modes. To address this issue, a hybrid method combining machine learning with equations of physical laws was employed for extrapolating the model. The dependencies obtained through this hybrid method exhibited an error margin of about 2%. The proposed method can be used to tackle several important problems in the digital transformation process of the Siberian Generating Company. By applying this method, the company can enhance its operational efficiency, particularly by compensating for measurement errors and ensuring compliance with material and energy balances. [ABSTRACT FROM AUTHOR]

The rapid growth of the Internet has led to an increased demand for online services. However, this surge in online activity has also brought about a new threat: phishing attacks. Phishing is a type of cyberattack that utilizes social engineering techniques and technological manipulations to steal crucial information from unsuspecting individuals. Consequently, there is a rising necessity to create dependable phishing URL detection models that can effectively identify phishing URLs with enhanced accuracy and reduced prediction overhead. This study introduces XAI-PhD, an innovative phishing detection method that utilizes machine learning (ML) and Shapley additive explanation (SHAP) capabilities. Specifically, XAI-PhD utilizes SHAP to thoroughly analyze the significance of each feature in influencing the decision-making process of the classifier. By selectively incorporating input characteristics based on their SHAP values, only the most crucial attributes are assessed, enabling the development of a highly adaptable and generalized model. XAI-PhD utilizes a lightweight gradient boosting machine as its classifier, and a series of rigorous tests are conducted to assess its performance compared to established baseline methods. The empirical findings unequivocally demonstrate the exceptional effectiveness of XAI-PhD, as evidenced by its remarkable accuracy and F1-score of 99.8% and 99%, respectively. Moreover, XAI-PhD exhibits high computational efficiency, requiring only 1.47 milliseconds and 18.5 microseconds per record to generate accurate predictions. [ABSTRACT FROM AUTHOR]

2D hybrid organic and inorganic perovskites (HOIPs) are used as capping layers on top of 3D perovskites to enhance their stability while maintaining the desired power conversion efficiency (PCE). Therefore, the 2D HOIP needs to withstand mechanical stresses and deformations, making the stiffness an important observable. However, there is no model for unravelling the relationship between their crystal structures and mechanical properties. In this work, explainable machine learning (ML) models are used to accelerate the in silico prediction of mechanical properties of 2D HOIPs, as indicated by their out‐of‐plane and in‐plane Young's modulus. The ML models can distinguish between stiff and non‐stiff 2D HOIPs, and extract the dominant physical feature influencing their Young's moduli, viz. the metal‐halogen‐metal bond angle. Furthermore, the steric effect index (STEI) of cations is found to be a rough criterion for non‐stiffness. Their optimal ranges are extracted from a probability analysis. Based on the strong correlation between the deformation of octahedra and the Young's modulus, the transferability of the approach from single‐layer to multi‐layer 2D HOIPs is demonstrated. This work represents a step toward unravelling the complex relationship between crystal structure and mechanical properties of 2D HOIPs using ML as a tool. [ABSTRACT FROM AUTHOR]

This study compares the calculation of snow water equivalent (SWE) using machine learning algorithms with the conventional degree-day method. The study uses machine learning techniques such as LASSO, Random Forest, Support Vector Machines, and CatBoost. It proposes an innovative use of feature engineering (FE) to improve the accuracy and robustness of SWE predictions by machine learning intended for interpolation, extrapolation, or imputation of missing data. The performance of machine learning approaches is evaluated against the traditional degree-day method for predicting SWE. The study emphasizes and demonstrates gains when modeling is enhanced by transforming basic, raw data through feature engineering. The results, verified in a case study from the mountainous region of Slovakia, suggest that machine learning, particularly CatBoost with feature engineering, shows better results in SWE estimation in comparison with the degree-day method, although the authors present a refined application of the degree-day method by utilizing genetic algorithms. Nevertheless, the study finds that the degree-day method achieved accuracy with a Nash–Sutcliffe coefficient of efficiency NSE = 0.59, while the CatBoost technique enhanced with the proposed FE achieved an accuracy NSE = 0.86. The results of this research contribute to refining snow hydrology modeling and optimizing SWE prediction for improved decision-making in snow-dominated regions. [ABSTRACT FROM AUTHOR]

In today's dynamic business world, organizations are increasingly relying on innovative technologies to improve the efficiency and effectiveness of their human resource (HR) management. Our study uses historical time and attendance data collected with the MojeUre time and attendance system to predict employee absenteeism, including sick and vacation leave, using machine learning methods. We integrate employee demographic data and the absence profiles on timesheets showing daily attendance patterns as fundamental elements for our analysis. We also convert the absence data into a feature-based format suitable for the machine learning methods used. Our primary goal in this paper is to evaluate how well we can predict sick leave and vacation leave over short- and long-term intervals using tree-based machine learning methods based on the predictive clustering paradigm. This paper compares the effectiveness of these methods in different learning settings and discusses their impact on improving HR decision-making processes. [ABSTRACT FROM AUTHOR]

With the proliferation of the Internet, network complexities for both commercial and state organizations have significantly increased, leading to more sophisticated and harder-to-detect network attacks. This evolution poses substantial challenges for intrusion detection systems, threatening the cybersecurity of organizations and national infrastructure alike. Although numerous deep learning techniques such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and graph neural networks (GNNs) have been applied to detect various network attacks, they face limitations due to the lack of standardized input data, affecting model accuracy and performance. This paper proposes a novel preprocessing method for flow data from network intrusion detection systems (NIDSs), enhancing the efficacy of a graph neural network model in malicious flow detection. Our approach initializes graph nodes with data derived from flow features and constructs graph edges through the analysis of IP relationships within the system. Additionally, we propose a new graph model based on the combination of the graph neural network (GCN) model and SAGEConv, a variant of the GraphSAGE model. The proposed model leverages the strengths while addressing the limitations encountered by the previous models. Evaluations on two IDS datasets, CICIDS-2017 and UNSW-NB15, demonstrate that our model outperforms existing methods, offering a significant advancement in the detection of network threats. This work not only addresses a critical gap in the standardization of input data for deep learning models in cybersecurity but also proposes a scalable solution for improving the intrusion detection accuracy. [ABSTRACT FROM AUTHOR]

Bearings are essential in machinery. Damage to them can cause financial losses and safety risks at industrial sites. Therefore, it is necessary to design an accurate diagnostic model. Although many bearing fault diagnosis methods have been proposed recently, they still cannot meet the requirements of high-accurate prediction of bearing faults. There are several challenges in this: 1) In practical settings, gathering sufficient and balanced sample data for training diagnostic network models proves challenging. 2) The damage to bearings in real industrial production sites is not singular, and compound faults are also a huge challenge for diagnostic networks. To address these issues, this study introduces a novel fault diagnosis model called EMALKNet that integrates DCGAN with Efficient Multi-Scale Attention (EMAGAN) and RepLKNet-XL, enhancing the detection and analysis of bearing faults in industrial machinery. This model employs EMAGAN to explore the underlying distribution of raw data, thereby enlarging the fault sample pool and enhancing the model's diagnostic capabilities; The large kernel structure of RepLKNet-XL is different from the current mainstream small kernel and has stronger representation extraction ability. The proposed method has been validated on the Paderborn University dataset and the Huazhong University of Science and Technology dataset. [ABSTRACT FROM AUTHOR]

The detection performance of radar is significantly impaired by active jamming and mutual interference from other radars. This paper proposes a radio signal modulation recognition method to accurately recognize these signals, which helps in the jamming cancellation decisions. Based on the ensemble learning stacking algorithm improved by meta-feature enhancement, the proposed method adopts random forests, K-nearest neighbors, and Gaussian naive Bayes as the base-learners, with logistic regression serving as the meta-learner. It takes the multi-domain features of signals as input, which include time-domain features including fuzzy entropy, slope entropy, and Hjorth parameters; frequency-domain features, including spectral entropy; and fractal-domain features, including fractal dimension. The simulation experiment, including seven common signal types of radar and active jamming, was performed for the effectiveness validation and performance evaluation. Results proved the proposed method's performance superiority to other classification methods, as well as its ability to meet the requirements of low signal-to-noise ratio and few-shot learning. [ABSTRACT FROM AUTHOR]

Automated fraud detection can assist organisations to safeguard user accounts, a task that is very challenging due to the great sparsity of known fraud transactions. Many approaches in the literature focus on credit card fraud and ignore the growing field of online banking. However, there is a lack of publicly available data for both. The lack of publicly available data hinders the progress of the field and limits the investigation of potential solutions. With this work, we: (a) introduce FraudNLP, the first anonymised, publicly available dataset for online fraud detection, (b) benchmark machine and deep learning methods with multiple evaluation measures, (c) argue that online actions do follow rules similar to natural language and hence can be approached successfully by natural language processing methods. [ABSTRACT FROM AUTHOR]

Fetal health care is vital in ensuring the health of pregnant women and the fetus. Regular check-ups need to be taken by the mother to determine the status of the fetus' growth and identify any potential problems. To know the status of the fetus, doctors monitor blood reports, Ultrasounds, cardiotocography (CTG) data, etc. Still, in this research, we have considered CTG data, which provides information on heart rate and uterine contractions during pregnancy. Several researchers have proposed various methods for classifying the status of fetus growth. Manual processing of CTG data is time-consuming and unreliable. So, automated tools should be used to classify fetal health. This study proposes a novel neural network-based architecture, the Dynamic Multi-Layer Perceptron model, evaluated from a single layer to several layers to classify fetal health. Various strategies were applied, including pre-processing data using techniques like Balancing, Scaling, Normalization hyperparameter tuning, batch normalization, early stopping, etc. to enhance the model's performance. A comparative analysis of the proposed method is done against the traditional machine learning models to showcase its accuracy (97%). An ablation study without any pre-processing techniques is also illustrated. This study easily provides valuable interpretations for healthcare professionals in the decision-making process. [ABSTRACT FROM AUTHOR]

Human Activity Recognition (HAR) using smartphone sensors has been identified as a significant emerging research domain. Its application areas exhibit the performance from the intelligent tailored activity monitoring. Researchers have proposed various HAR models to recognize the human activity patterns using traditional smartphone sensor data. In addition, embedding contextual information such as data availability, sensing device orientation, body part location, axis layout, and more makes a fruitful impact on the quality of activity sensor data. The research challenge occurred due to the lack of contextual information in sensor data, leading to activity patterns ambiguity. Often, the motion sensor was separately used to acquire contextual information that consumes unnecessary computational resources. In this paper, we have used activity sensor data availability as contextual information. The proposed Deep Context Model (DCM) recognizes the activity pattern in the dual context-attention mode, i.e., static and dynamic context. The proposed model consists of convolutional and recurrent networks that find the associated activity patterns in a dual context. The convolutional networks are excellent for automatic feature extraction in a static context, whereas recurrent networks are used in a dynamic context for memorizing patterns. We evaluated the performance of the proposed model using the open accessible KU-HAR dataset. The experimental outcomes revealed that DCM has achieved the F1 score of 98.96% and 99.62% for static and dynamic context, respectively. Further, the robustness and applicability of the proposed model have been gauged using the HHAR dataset. [ABSTRACT FROM AUTHOR]

Lexical answer type prediction is integral to biomedical question–answering systems. LAT prediction aims to predict the expected answer's semantic type of a factoid or list-type biomedical question. It also aids in the answer processing stage of a QA system to assign a high score to the most relevant answers. Although considerable research efforts exist for LAT prediction in diverse domains, it remains a challenging biomedical problem. LAT prediction for the biomedical field is a multi-label classification problem, as one biomedical question might have more than one expected answer type. Achieving high performance on this task is challenging as biomedical questions have limited lexical features. One biomedical question must be assigned multiple labels given these limited lexical features. In this paper, we develop a novel feature set (lexical, noun concepts, verb concepts, protein–protein interactions, and biomedical entities) from these lexical features. Using ensemble learning with bagging, we use the label power set transformation technique to classify multi-label. We evaluate the integrity of our proposed methodology on the publicly available multi-label biomedical questions dataset (MLBioMedLAT) and compare it with twelve state-of-the-art multi-label classification algorithms. Our proposed method attains a micro-F1 score of 77%, outperforming the baseline model by 25.5%. [ABSTRACT FROM AUTHOR]

The creation of features makes the difference in improving the photovoltaic forecast (PVF) for on‐grid, hybrid and off‐grid PV systems. The importance of the PVF is tremendous, and it can be essential in optimizing the home appliances to maximize the Renewable Energy Sources (RES) usage or to create performant bids for the electricity market. Several use cases are considered from the connectivity point of view. Therefore, in this paper, we propose a Weather Prediction Error (WPE)‐based method that uses a Stacking Regressor (SR) for various PV systems that coexist in the emerging Energy Communities (EC) landscape. The novelty of the research we conduct consists in proposing several features and determining the coefficients to adjust the PVF based on WPE. The forecast results of four types of PV systems from size and connectivity point of view are investigated. Compared with individual Machine Learning (ML) models, R2 increases with more than 3% with the SR and with more than 6% after applying the adjustment coefficients. Nevertheless, the major improvement is recorded for the off‐grid inverter and for the large industrial PV power plant, demonstrating that the proposed model is suitable for these types of systems. The other metrics improved as well, especially Mean Average Error (MAE) that decreases between 10% and 23%. A significant decrease is in the case of the industrial on‐grid PV, from 130 kW to 123 kW using the SR and to 107 kW after adjustments. This represents around 18% from the initial MAE. The ratio of daily deviations is also improved using the SR. For all the PV systems, the ratio stabilizes in a shorter interval, from daily values between 0.78 and 1.33 obtained with the ML models to values between 0.83 and 1.24 obtained with the SR. After the final adjustments, the interval becomes shorter, having daily values between 0.90 and 1.10. [ABSTRACT FROM AUTHOR]

This work describes a mathematical model for handwriting devices without a specific reference surface (SRS). The research was carried out on two hypotheses: the first considers possible circular segments that could be made during execution for the reconstruction of the trace, and the second is the combination of lines and circles. The proposed system has no flat reference surface, since the sensor is inside the pencil that describes the trace, not on the surface as in tablets or cell phones. An inertial sensor was used for the measurements, in this case, a commercial Micro-Electro Mechanical sensor of linear acceleration. The tracking device is an IMU sensor and a processing card that allows inertial measurements of the pen during on-the-fly tracing. It is essential to highlight that the system has a non-inertial reference frame. Comparing the two proposed models shows that it is possible to construct shapes from curved lines and that the patterns obtained are similar to what is recognized; this method provides an alternative to quaternion calculus for poorly specified orientation problems. [ABSTRACT FROM AUTHOR]

Abstract Breast cancer (BC) is a prominent cause of female mortality on a global scale. Recently, there has been growing interest in utilizing blood and tissue-based biomarkers to detect and diagnose BC, as this method offers a non-invasive approach. To improve the classification and prediction of BC using large biomarker datasets, several machine-learning techniques have been proposed. In this paper, we present a multi-stage approach that consists of computing new features and then sorting them into an input image for the ResNet50 neural network. The method involves transforming the original values into normalized values based on their membership in the Gaussian distribution of healthy and BC samples of each feature. To test the effectiveness of our proposed approach, we employed the Coimbra and Wisconsin datasets. The results demonstrate efficient performance improvement, with an accuracy of 100% and 100% using the Coimbra and Wisconsin datasets, respectively. Furthermore, the comparison with existing literature validates the reliability and effectiveness of our methodology, where the normalized value can reduce the misclassified samples of ML techniques because of its generality.

In the realm of data science, dealing with real-world datasets often presents a formidable challenge, primarily due to the sheer volume of features that significantly lack relevance or may be redundant. Effective feature engineering is vital in constructing robust ensemble ML models, where the choice of input features influences overall performance. Towards this, the present research presents a novel framework to feature engineering by hybridizing the MRMR insights and Chi2 independence techniques. MRMR emphasizes feature relevance and non-redundancy, while Chi2 quantifies the independence of features from the target variable. The hybrid framework adheres to the incremental feature engineering approach, with the goal of improving predictive accuracy, model robustness, and adaptability. Through extensive experimentation on employed water quality dataset, the framework illustrates the superiority of hybrid model over using MRMR and Chi2 independently. The results of the proposed HFE-EML exhibit substantial improvements, reaching approximately 99.10% in ensemble machine learning models' performance, reduced overfitting, and enhanced generalization.

Purpose – Surface roughness has a serious impact on the fatigue strength, wear resistance and life of mechanical products. Realizing the evolution of surface quality through theoretical modeling takes a lot of effort. To predict the surface roughness of milling processing, this paper aims to construct a neural network based on deep learning and data augmentation. Design/methodology/approach – This study proposes a method consisting of three steps. Firstly, the machine tool multisource data acquisition platform is established, which combines sensor monitoring with machine tool communication to collect processing signals. Secondly, the feature parameters are extracted to reduce the interference and improve the model generalization ability. Thirdly, for different expectations, the parameters of the deep belief network (DBN) model are optimized by the tent-SSA algorithm to achieve more accurate roughness classification and regression prediction. Findings – The adaptive synthetic sampling (ADASYN) algorithm can improve the classification prediction accuracy of DBN from 80.67% to 94.23%. After the DBN parameters were optimized by Tent-SSA, the roughness prediction accuracy was significantly improved. For the classification model, the prediction accuracy is improved by 5.77% based on ADASYN optimization. For regression models, different objective functions can be set according to production requirements, such as root-mean-square error (RMSE) or MaxAE, and the error is reduced by more than 40% compared to the original model. Originality/value – A roughness prediction model based on multiple monitoring signals is proposed, which reduces the dependence on the acquisition of environmental variables and enhances the model's applicability. Furthermore, with the ADASYN algorithm, the Tent-SSA intelligent optimization algorithm is introduced to optimize the hyperparameters of the DBN model and improve the optimization performance.

Abstract Background Single-cell chromatin accessibility assays, such as scATAC-seq, are increasingly employed in individual and joint multi-omic profiling of single cells. As the accumulation of scATAC-seq and multi-omics datasets continue, challenges in analyzing such sparse, noisy, and high-dimensional data become pressing. Specifically, one challenge relates to optimizing the processing of chromatin-level measurements and efficiently extracting information to discern cellular heterogeneity. This is of critical importance, since the identification of cell types is a fundamental step in current single-cell data analysis practices. Results We benchmark 8 feature engineering pipelines derived from 5 recent methods to assess their ability to discover and discriminate cell types. By using 10 metrics calculated at the cell embedding, shared nearest neighbor graph, or partition levels, we evaluate the performance of each method at different data processing stages. This comprehensive approach allows us to thoroughly understand the strengths and weaknesses of each method and the influence of parameter selection. Conclusions Our analysis provides guidelines for choosing analysis methods for different datasets. Overall, feature aggregation, SnapATAC, and SnapATAC2 outperform latent semantic indexing-based methods. For datasets with complex cell-type structures, SnapATAC and SnapATAC2 are preferred. With large datasets, SnapATAC2 and ArchR are most scalable.

This study investigates the utilization of artificial intelligence (AI) for feature engineering in the education sector, highlighting its potential to enhance individualized learning and improve academic outcomes. The correlation analysis, performed using a correlation matrix of the feature set, indicated that specific pairings of characteristics exhibit a strong association, resulting in the ineffectiveness of conventional models. In order to tackle this issue, we utilized three sophisticated machine learning methodologies: Adaptive Lasso (ALasso), Artificial Neural Networks (ANN), and Support Vector Regression (SVR). The ALasso model discovered several influential characteristics, namely Gender (X5), Education (X1), Hours of Work (X4), and Marital Status (X6), that significantly affect salaries. Subsequently, a comparative evaluation of these methods was conducted using Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). The results demonstrated that SVR outperformed the other techniques, with the most optimal RMSE of 0.595 and MAE of 0.423. These findings emphasize the significance of using data-driven strategies in policymaking and propose further investigation into the use of AI methods in various educational contexts to improve the identification of features and the performance of models.

Urban water demand prediction is not only the foundation of water resource planning and management, but also an important component of water supply system optimization and scheduling. Therefore, predicting future water demand is of great significance. For univariate time series data, the issue of outliers can be solved through data preprocessing. Then, the data input dimension is increased through feature engineering, and finally, the LightGBM (Light Gradient Boosting Machine) model is used to predict future water demand. The results demonstrate that cubic polynomial interpolation outperforms the Prophet model and the linear method in the context of missing value interpolation tasks. In terms of predicting water demand, the LightGBM model demonstrates excellent forecasting performance and can effectively predict future water demand trends. The evaluation indicators MAPE (mean absolute percentage error) and NSE (Nash–Sutcliffe efficiency coefficient) on the test dataset are 4.28% and 0.94, respectively. These indicators can provide a scientific basis for short-term prediction of water supply enterprises. HIGHLIGHTS Interpolation of raw training data may not necessarily improve the performance of predictive models.; Accurate prediction of univariate data can be achieved through feature engineering and machine learning.;

Abstract Detecting spam emails is essential to maintaining the security and integrity of email communication. Existing research has made significant progress in developing effective spam detection models, but challenges remain in improving classification performance and adaptability to evolving spamming techniques. In this study, we propose a novel spam detection model with a comprehensive feature engineering approach that combines term frequency-inverse document frequency (TF-IDF) vectorizer and word embedding features to optimize the feature space. Our contribution lies in integrating semantic-based word embeddings, leveraging pre-existing knowledge to capture the semantic meaning of words and enhance the representation of email texts. To identify the most suitable word embedding technique for our model, we evaluated GloVe, Word2Vec, and FastText. GloVe was selected for its better performance, which is the result of its pre-training on a large and diverse text corpus. Furthermore, the model was evaluated without word embeddings, which did not exhibit the same effectiveness level as our word embedding-based model. Additionally, we utilized the support vector machine as a classifier and hyperparameter tuning technique to identify our model’s most effective parameter values. The proposed model was tested on two datasets. The experimental results showed that our model outperformed the other models discussed in the literature, achieving an accuracy of 99.5% on the SpamAssassin dataset, and 99.28% on the Enron-Spam dataset.

Abstract With the increase in the generation and spread of textual content on social media, natural language processing (NLP) has become an important area of research for detecting underlying threats, racial abuse, violence, and implied warnings in the content. The subtlety and ambiguity of language make the development of effective models for detecting threats in text a challenging task. This task is further complicated when the threat is not explicitly conveyed. This study focuses on the task of implied threat detection using an explicitly designed machine-generated dataset with both linguistic and lexical features. We evaluated the performance of different machine learning algorithms on these features including Support Vector Machines, Logistic Regression, Naive Bayes, Decision Tree, and K-nearest neighbors. The ensembling approaches of Adaboost, Random Forest, and Gradient Boosting were also explored. Deep learning modeling was performed using Long Short-Term Memory, Deep Neural Networks (DNN), and Bidirectional Long Short-Term Memory (BiLSTM). Based on the evaluation, it was observed that classical and ensemble models overfit while working with linguistic features. The performance of these models improved when working with lexical features. The model based on logistic regression exhibited superior performance with an F 1 score of 77.13%. While experimenting with deep learning models, DNN achieved an F 1 score of 91.49% while the BiLSTM achieved an F 1 score of 91.61% while working with lexical features. The current study provides a baseline for future research in the domain of implied threat detection.

Abstract Background Machine learning (ML) has emerged as the predominant computational paradigm for analyzing large-scale datasets across diverse domains. The assessment of dataset quality stands as a pivotal precursor to the successful deployment of ML models. In this study, we introduce DREAMER (Data REAdiness for MachinE learning Research), an algorithmic framework leveraging supervised and unsupervised machine learning techniques to autonomously evaluate the suitability of tabular datasets for ML model development. DREAMER is openly accessible as a tool on GitHub and Docker, facilitating its adoption and further refinement within the research community.. Results The proposed model in this study was applied to three distinct tabular datasets, resulting in notable enhancements in their quality with respect to readiness for ML tasks, as assessed through established data quality metrics. Our findings demonstrate the efficacy of the framework in substantially augmenting the original dataset quality, achieved through the elimination of extraneous features and rows. This refinement yielded improved accuracy across both supervised and unsupervised learning methodologies. Conclusion Our software presents an automated framework for data readiness, aimed at enhancing the integrity of raw datasets to facilitate robust utilization within ML pipelines. Through our proposed framework, we streamline the original dataset, resulting in enhanced accuracy and efficiency within the associated ML algorithms.

Diabetes Mellitus (DM) is an enduring metabolic illness that disturbs many individuals globally. This study addresses the global impact of Diabetes Mellitus (DM) and emphasizes the critical role of accurate DM detection in early diagnosis, effective treatment, and prevention of complications. The research introduces an optimized DM detection model, the GBM-DRU (Gradient Boosting Machine - Data Reduction Unit) network, which integrates feature engineering and ensemble learning techniques to enhance prediction accuracy and support clinical decision-making. The GBM-DRU network combines the powerful gradient boosting machine algorithm with a data reduction unit (DRU) for efficient feature selection, reducing dimensionality and improving computational efficiency. Feature engineering enhances discriminatory power, while ensemble learning methods, including bagging and boosting, improve overall model performance. Rigorous experiments on a comprehensive dataset of DM patients demonstrate that the proposed approach outperforms existing models in terms of accuracy, sensitivity, specificity, and AUC-ROC. The optimized model provides valuable insights into feature importance, aiding clinical decision-making and deepening the understanding of DM risk factors. Therefore, the GBM-DRU network, utilizing feature engineering and ensemble learning, presents a viable approach to precise diagnosis of diabetes mellitus, with favorable implications for patient outcomes, disease control, and public health campaigns. The improved prediction accuracy, feature interpretability, and clinical decision support capabilities of the model may have a beneficial effect on public health campaigns, disease management, and patient outcomes.

Hyperspectral imaging holds significant promise in remote sensing applications, particularly for land cover and land-use classification, thanks to its ability to capture rich spectral information. However, leveraging hyperspectral data for accurate segmentation poses critical challenges, including the curse of dimensionality and the scarcity of ground truth data, that hinder the accuracy and efficiency of machine learning approaches. This paper presents a holistic approach for adaptive optimized hyperspectral-based land cover and land-use segmentation using automated machine learning (AutoML). We address the challenges of high-dimensional hyperspectral data through a revamped machine learning pipeline, thus emphasizing feature engineering tailored to hyperspectral classification tasks. We propose a framework that dissects feature engineering into distinct steps, thus allowing for comprehensive model generation and optimization. This framework incorporates AutoML techniques to streamline model selection, hyperparameter tuning, and data versioning, thus ensuring robust and reliable segmentation results. Our empirical investigation demonstrates the efficacy of our approach in automating feature engineering and optimizing model performance, even without extensive ground truth data. By integrating automatic optimization strategies into the segmentation workflow, our approach offers a systematic, efficient, and scalable solution for hyperspectral-based land cover and land-use classification. [ABSTRACT FROM AUTHOR]

Detecting spam emails is essential to maintaining the security and integrity of email communication. Existing research has made significant progress in developing effective spam detection models, but challenges remain in improving classification performance and adaptability to evolving spamming techniques. In this study, we propose a novel spam detection model with a comprehensive feature engineering approach that combines term frequency-inverse document frequency (TF-IDF) vectorizer and word embedding features to optimize the feature space. Our contribution lies in integrating semantic-based word embeddings, leveraging pre-existing knowledge to capture the semantic meaning of words and enhance the representation of email texts. To identify the most suitable word embedding technique for our model, we evaluated GloVe, Word2Vec, and FastText. GloVe was selected for its better performance, which is the result of its pre-training on a large and diverse text corpus. Furthermore, the model was evaluated without word embeddings, which did not exhibit the same effectiveness level as our word embedding-based model. Additionally, we utilized the support vector machine as a classifier and hyperparameter tuning technique to identify our model's most effective parameter values. The proposed model was tested on two datasets. The experimental results showed that our model outperformed the other models discussed in the literature, achieving an accuracy of 99.5% on the SpamAssassin dataset, and 99.28% on the Enron-Spam dataset. [ABSTRACT FROM AUTHOR]

A key obstacle in personalised fashion recommendations is the challenge of capturing user physical attributes at a large scale, which limits exclusively computational methods (like machine learning) to readily available attributes whose influence on recommendation accuracy is variable. Expert advice is a potential means of identifying influential user attributes. However, individual experts often disagree or offer conflicting advice. Thus, identifying areas where expert advice is or isn't consistent, in the context of user attributes and profiling is critical. Here, we characterise the breadth of expert definitions of user attributes and profiles through an exhaustive assessment of 156 years of advice literature. Expert definitions of body colouring, shape, and personality attributes are extracted and compared. The range of attribute-value relationships and profile definitions in each domain is described, and coherence among authors for each domain is discussed. [ABSTRACT FROM AUTHOR]

General physicians misclassify depression in more than half of the cases. Researchers have explored the feasibility of leveraging passively collected data points, also called digital biomarkers, to provide more granular understanding of depression phenotypes as well as a more objective assessment of disease. This paper provides a systematic review following the PRISMA guidelines (Page et al., 2021) to understand which digital biomarkers might be relevant for passive screening of depression. Pubmed and PsycInfo were systematically searched for studies published from 2019 to early 2024, resulting in 161 records assessed for eligibility. Excluded were intervention studies, studies focusing on a different disease or those with a lack of passive data collection. 74 studies remained for a quality assessment, after which 27 studies were included. The review shows that depressed participants' real-life behavior such as reduced communication with others can be tracked by passive data. Machine learning models for the classification of depression have shown accuracies up to 0.98, surpassing the quality of many standardized assessment methods. Inconsistency of outcome reporting of current studies does not allow for drawing statistical conclusions regarding effectiveness of individual included features. The Covid-19 pandemic might have impacted the ongoing studies between 2020 and 2022. While digital biomarkers allow real-life tracking of participant's behavior and symptoms, further work is required to align the feature engineering of digital biomarkers. With shown high accuracies of assessments, connecting digital biomarkers with clinical practice can be a promising method of detecting symptoms of depression automatically. • This paper provides a PRISMA-guided systematic review of studies (2019-2024), showing model accuracy ranges of 0.52-0.98. • The review summarises constructs related to mood, location, smartphone use, communication, sleep quality and heart rate. • To be integrated into clinical practice, captured data needs to be reduced to the most relevant features. [ABSTRACT FROM AUTHOR]

Online platforms have become the primary means for travellers to search, compare, and book accommodations for their trips. Consequently, online platforms and revenue managers must acquire a comprehensive comprehension of these dynamics to formulate a competitive and appealing offerings. Recent advancements in natural language processing, specifically through the development of large language models, have demonstrated significant progress in capturing the intricate nuances of human language. On the other hand, knowledge graphs have emerged as potent instruments for representing and organizing structured information. Nevertheless, effectively integrating these two powerful technologies remains an ongoing challenge. This paper presents an innovative deep learning methodology that combines large language models with domain-specific knowledge graphs for classification of tourism offers. The main objective of our system is to assist revenue managers in the following two fundamental dimensions: (i) comprehending the market positioning of their accommodation offerings, taking into consideration factors such as accommodation price and availability, together with user reviews and demand, and (ii) optimizing presentations and characteristics of the offerings themselves, with the intention of improving their overall appeal. For this purpose, we developed a domain knowledge graph covering a variety of information about accommodations and implemented targeted feature engineering techniques to enhance the information representation within a large language model. To evaluate the effectiveness of our approach, we conducted a comparative analysis against alternative methods on four datasets about accommodation offers in London. The proposed solution obtained excellent results, significantly outperforming alternative methods. [ABSTRACT FROM AUTHOR]

The proliferation of data collection technologies often results in large data sets with many observations and many variables. In practice, highly relevant engineered features are often groups of predictors that share a common regression coefficient (i.e., the predictors in the group affect the response only via their collective sum), where the groups are unknown in advance and must be discovered from the data. We propose an algorithm called coefficient tree regression (CTR) to discover the group structure and fit the resulting regression model. In this regard CTR is an automated way of engineering new features, each of which is the collective sum of the predictors within each group. The algorithm can be used when the number of variables is larger than, or smaller than, the number of observations. Creating new features that affect the response in a similar manner improves predictive modeling, especially in domains where the relationships between predictors are not known a priori. CTR borrows computational strategies from both linear regression (fast model updating when adding/modifying a feature in the model) and regression trees (fast partitioning to form and split groups) to achieve outstanding computational and predictive performance. Finding features that represent hidden groups of predictors (i.e., a hidden ontology) that impact the response only via their sum also has major interpretability advantages, which we demonstrate with a real data example of predicting political affiliations with television viewing habits. In numerical comparisons over a variety of examples, we demonstrate that both computational expense and predictive performance are far superior to existing methods that create features as groups of predictors. Moreover, CTR has overall predictive performance that is comparable to or slightly better than the regular lasso method, which we include as a reference benchmark for comparison even though it is non-group-based, in addition to having substantial computational and interpretive advantages over lasso. [ABSTRACT FROM AUTHOR]

Time series forecasting is a challenging task with applications in a wide range of domains. Auto-regression is one of the most common approaches to address these problems. Accordingly, observations are modelled by multiple regression using their past lags as predictor variables. We investigate the extension of auto-regressive processes using statistics which summarise the recent past dynamics of time series. The result of our research is a novel framework called VEST, designed to perform feature engineering using univariate and numeric time series automatically. The proposed approach works in three main steps. First, recent observations are mapped onto different representations. Second, each representation is summarised by statistical functions. Finally, a filter is applied for feature selection. We discovered that combining the features generated by VEST with auto-regression significantly improves forecasting performance in a database composed by 90 time series with high sampling frequency. However, we also found that there are no improvements when the framework is applied for multi-step forecasting or in time series with low sample size. VEST is publicly available online. [ABSTRACT FROM AUTHOR]

Deformation effectively represents the structural integrity of concrete dams and acts as a clear indicator of their operational performance. Predicting deformation is critical for monitoring the safety of hydraulic structures. To this end, this paper proposes an artificial intelligence-based process for predicting concrete dam deformation. Initially, using the principles of feature engineering, the preprocessing of deformation safety monitoring data is conducted. Subsequently, employing a stacking model fusion method, a novel prediction process embedded with multiple artificial intelligence algorithms is developed. Moreover, three new performance indicators—a superiority evaluation indicator, an accuracy evaluation indicator, and a generalization evaluation indicator—are introduced to provide a comprehensive assessment of the model's effectiveness. Finally, an engineering example demonstrates that the ensemble artificial intelligence method proposed herein outperforms traditional statistical models and single machine learning models in both fitting and predictive accuracy, thereby providing a scientific and effective foundation for concrete dam deformation prediction and safety monitoring. [ABSTRACT FROM AUTHOR]