Feature Creation Research Articles

ANALYSING MARKET DYNAMICS: REVEALING OBSCURED PATTERNS IN LQ45 STOCKS (2021-2023) USING WARD’S HIERARCHICAL CLUSTERING

This study aimed to address the instability of the Indonesian stock market from 2021 to 2023 by analyzing the LQ45 index, a critical indicator of economic robustness and corporate performance. Hierarchical Ward clustering was employed to categorize LQ45 stocks based on fundamental metrics such as Return, Volume, Price, Price-Earnings Ratio (PER), Earnings Per Share (EPS), and Dividends. Data preprocessing involved feature creation, Max-Abs scaling for normalization, and binary encoding of categorical variables. The optimal number of clusters was identified using dendrograms, revealing two primary clusters: one focusing on core materials and the other on financial services, alongside other industry-specific clusters. This method, characterized by its ability to minimize variance within clusters and determine natural groupings without predefined assumptions, provided valuable insights for financial advisors, policymakers, and investors. The findings offer practical guidance for optimizing decision-making, minimizing risks, and leveraging opportunities within the Indonesian stock market during a period of significant economic uncertainty. By employing this strategy, investors and traders can gain a comprehensive understanding of the current condition of the stock market, offering a thorough comprehension of the connections between equities and the operational and financial issues currently under scrutiny.

Multilevel hybrid handcrafted feature extraction based depression recognition method using speech

Background and purposeDiagnosis of depression is based on tests performed by psychiatrists and information provided by patients or their relatives. In the field of machine learning (ML), numerous models have been devised to detect depression automatically through the analysis of speech audio signals. While deep learning approaches often achieve superior classification accuracy, they are notably resource-intensive. This research introduces an innovative, multilevel hybrid feature extraction-based classification model, specifically designed for depression detection, which exhibits reduced time complexity. Materials and methodsMODMA dataset consisting of 29 healthy and 23 Major depressive disorder audio signals was used. The constructed model architecture integrates multilevel hybrid feature extraction, iterative feature selection, and classification processes. During the Hybrid Handcrafted Feature (HHF) generation stage, a combination of textural and statistical methods was employed to extract low-level features from speech audio signals. To enhance this process for high-level feature creation, a Multilevel Discrete Wavelet Transform (MDWT) was applied. This technique produced wavelet subbands, which were then input into the hybrid feature extractor, enabling the extraction of both high and low-level features. For the selection of the most pertinent features from these extracted vectors, Iterative Neighborhood Component Analysis (INCA) was utilized. Finally, in the classification phase, a one-dimensional nearest neighbor classifier, augmented with ten-fold cross-validation, was implemented to achieve detailed, results. ResultsThe HHF-based speech audio signal classification model attained excellent performance, with the 94.63 % classification accuracy. ConclusionsThe findings validate the remarkable proficiency of the introduced HHF-based model in depression classification, underscoring its computational efficiency.

BRAVE: A cascaded generative model with sample attention for robust few shot image classification

Few-shot learning (FSL) confronts notable challenges due to the disparity between training and testing categories, leading to channel bias in neural networks and hindering accurate feature discernment. To address this, we introduce Biased-Reduction Attentive Network (BRAVE), an innovative model that incorporates a refined Vector Quantized Variational Autoencoder (VQ-VAE) backbone, enhanced with our Diverse Quantization (DQ) Module, for unbiased, fine-grained feature creation. Alongside, our Sample Attention (SA) Module is utilized for extracting discriminative features from these unbiased, fine-grained features. The DQ Module in BRAVE strategically integrates prior distribution regularization and stochastic masking with Gumbel sampling for balanced and diverse codebook engagement, while the SA Module leverages inter-sample dynamics for identifying critical features. This synergy effectively counters channel bias and boosts classification accuracy in FSL setups, surpassing current leading methods. Our approach represents a practical balance between preserving detailed features through the decoder and ensuring classification effectiveness, marking a significant advance in FSL. BRAVE’s implementation is accessible for community use and further exploration. Code and models available at https://github.com/ApocalypsezZ/BRAVE.

Machine-learning-based sampling inspection under OQC capacity for real-time quality monitoring in the TFT-LCD industry

ABSTRACT In the thin film transistor-liquid crystal display (TFT-LCD) industry, the competition is fierce, and the external failure cost is high. To improve the quality of released products, sampling inspection under outgoing quality control (OQC) is conducted before products are released. Owing to the variety and complexity of inspection items, the task must be carried out manually while realising automation is difficult. Traditional random sampling inspection is limited by sampling quantity, manpower, and material resources. To effectively improve acceptable quality levels, this paper proposes a quality predictive monitoring framework tailored to the production and inspection characteristics of the TFT-LCD industry. This framework includes real-time automatic quality monitoring based on an optimal machine learning prediction model with two-phase feature creation, and it also addresses OQC inspection capacity constraints. Experimental results show the prediction performance based on proposed model is better than the traditional random sampling process. In practice, superior average outgoing quality improves by more than 70% at the same sampling level. In addition, the features of the process can be further controlled and managed based on the prediction model to monitor the key parameters in real time and respond to abnormalities quickly, thereby further improving the quality of released products.

Using ensemble learning and hierarchical strategy to predict the outcomes of ESWL for upper ureteral stone treatment

Urinary tract stones are a common and frequently recurring medical issue. Accurately predicting the success rate after surgery can help avoid ineffective medical procedures and reduce unnecessary healthcare costs. This study collected data from patients with upper ureter stones who underwent extracorporeal shock wave lithotripsy, including cases of successful as well as unsuccessful stone removal after the first and second lithotripsy procedures, and constructed prediction systems for the outcomes of the first and second lithotripsy procedures. Features were extracted from three categories of information: patient characteristics, stone characteristics, and extracorporeal shock wave lithotripsy machine data, and additional features were created using Feature Creation. Finally, the impact of features on the models was analyzed using six methods to calculate feature importance. Our prediction model for the first lithotripsy, selected from among 43 methods and seven ensemble learning techniques, achieves an AUC of 0.91. For the second lithotripsy, the AUC reaches 0.76. The results indicate that the detailed and binary information provided by patients regarding their history of stone experiences contributes differently to the predictive accuracy of the first and second lithotripsy procedures. The prediction tool is available at https://predictor.isu.edu.tw/ks.

Global Mean Based nearest Feature object Value Selection with Feature creation Method for Clustering Accuracy Improvement

Global Mean Based nearest Feature object Value Selection with Feature creation Method for Clustering Accuracy Improvement

Fault Prognosis in Smart Distribution System with Distributed Generation Using Fast Fourier Transform Assisted Machine Learning

The affordability and accessibility of electrical data related to microgrids and active distribution networks (ADN) have significantly improved due to smart digital relays with advanced communication capabilities. These relays can communicate with each other and central protection systems. Due to data accessibility a machine learning (ML) based module is proposed. This module predicts both fault type and location within an IEEE 34 bus active distribution system featuring distributed generations. To design this module, current measurements are retrieved from five specified bus locations and further Fast Fourier Transform is applied for feature extraction. The feature creation method is much simpler without any complexity. These features are further used for ML module’s training. Practical ADN operating conditions are considered when curating the training dataset. Weighted K-nearest neighbors are used for fault type prediction and linear regression is applied for fault location estimation and they are proven to be superior to other ML approaches. The obtained results show high accuracy and reliability in predictions, making these modules suitable for ADN protection. They also exhibit robustness against input measurement noise, ensuring consistent protection in various operating scenarios. The ADN is simulated in SIMULINK 2020a, and ML modeling is done using MATLAB 2020a.

The Role of Business Management in the Innovative Development of the Cultural Tourism and Education Industry

The Role of Business Management in the Innovative Development of the Cultural Tourism and Education Industry

High-throughput discovery of metal oxides with high thermoelectric performance via interpretable feature engineering on small data

In this work, we have proposed a data-driven screening framework combining the interpretable machine learning with high-throughput calculations to identify a series of metal oxides that exhibit both high-temperature tolerance and high power factors. Aiming at the problem of weak generalization ability of small data with power factors at high temperatures, we employ symbolic regression for feature creation which enhances the robustness of the model while preserving the physical meaning of features. 33 candidate metal oxides are finally targeted for high-temperature thermoelectric applications from a pool of 48,694 compounds in the Materials Project database. The Boltzmann transport theory is utilized to perform electrical transport properties calculations at 1,000 K. The relaxation time is approximated by employing constant electron-phonon coupling based on the deformation potential theory. Considering band degeneracy, the electron group velocity is obtained using the momentum matrix element method, yielding 28 materials with power factors greater than 50 μWcm−1K−2. The high-throughput framework we proposed is instrumental in the selection of metal oxides for high-temperature thermoelectric applications. Furthermore, our data-driven analysis and transport calculation suggest that metal oxides rich in elements such as cerium (Ce), tin (Sn), and lead (Pb) tend to exhibit high power factors at high temperatures.

Cartographic Resources for Equitable University–Community Interaction in Slum Areas

Cartographic resources play a crucial role in facilitating communication across various sectors, including research projects focused on low-income communities. Despite this, some researchers still adhere to colonialist and exploitative approaches. This study aims to promote equitable university–community interaction though cartographic resources, aid academic and vulnerable community users in choosing a better platform for their work, and provide insights to developers for improving the platforms to better serve the user profiles of community members. To achieve this, we examined the use of cartographic resources in five projects within low-income communities (commonly referred to as favelas or so-called “slums”) in three Brazilian cities, all guided by equitable principles. The study unfolds in four stages: (i) data collection from documents and interviews; (ii) systematization into seven analytical categories—cartographic resources, data, personnel, processes, equipment, general objectives, and specific objectives; (iii) analysis of eight cartographic resources; and (iv) a critical examination of the outcomes. The synthesis of the collected information identified 65 characteristics/demands, with 17 common to all projects, including vector feature creation, thematic map design, printed map usage, and satellite imagery. We also identified 53 geographic information system (GIS) functionalities required for the projects, predominantly related to vector data generation and editing. The outcomes demonstrate the benefits of project methodologies, contributing to a decolonial university–community praxis. Additionally, they underscore the potential of digital cartographic resources, functioning not solely as data collection tools but also as powerful instruments that empower slum residents to advocate for improvements and foster local development.

Optimizing Mental Health Detection in Indian Armed Forces Personnel through Feature Engineering Driven Dataset Reduction, Addressing Suicide, Depression, and Stress

Within the realm of machine learning, the construction of high-quality datasets stands as a crucial factor profoundly influencing model performance. This research aims to furnish a comprehensive guide for enhancing the accuracy and efficiency of dataset construction. It achieves this by integrating multi-variate reduction techniques and innovative feature engineering strategies, implemented within the Python programming ecosystem. As the landscape of datasets becomes increasingly diverse and complex, the imperative to optimize precision grows more critical. This study explores the judicious application of dimensionality reduction methods, such as Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-SNE), alongside various feature selection approaches to strategically streamline datasets while preserving vital information. In conjunction with these reduction techniques, the research introduces novel feature engineering methods to amplify the discriminative power of remaining features, thereby enriching the dataset's representational capacity. The exploration spans a spectrum of multi-variate reduction techniques and delves into feature engineering methodologies, including polynomial feature creation, interaction term generation, and domain-specific transformation functions. Practical implementations of these techniques are demonstrated through Python, showcasing their applicability across diverse domains. Empirical evaluations on real-world datasets underscore the efficacy of the proposed methodology, revealing superior accuracy and efficiency compared to conventional dataset construction approaches. The insights derived from this research contribute significantly to the broader discourse in machine learning, presenting a generic yet potent framework for enhancing precision in datasets. Beyond deepening our understanding of multi-variate reduction and feature engineering, the findings offer a practical guide for researchers and practitioners seeking to optimize precision in various machine learning applications.

Feature Creation to Enhance Explainability and Predictability of ML Models Using XAI

Feature Creation to Enhance Explainability and Predictability of ML Models Using XAI

Constructing Features for Screening Neurodevelopmental Disorders Using Grammatical Evolution

Developmental domains refer to different areas of a child’s growth and maturation, including physical, language, cognitive, and social–emotional skills. Understanding these domains helps parents, caregivers, and professionals track a child’s progress and identify potential areas of concern. Nevertheless, due to the high level of heterogeneity and overlap, neurodevelopmental disorders may go undiagnosed in children for a crucial period. Detecting neurodevelopmental disorders at an early stage is fundamental. Digital tools like artificial intelligence can help clinicians with the early detection process. To achieve this, a new method has been proposed that creates artificial features from the original ones derived from the SmartSpeech project, using a feature construction procedure guided by the Grammatical Evolution technique. The new features from a machine learning model are used to predict neurodevelopmental disorders. Comparative experiments demonstrated that using the feature creation method outperformed other machine learning methods for predicting neurodevelopmental disorders. In many cases, the reduction in the test error reaches up to 65% to the next better one.

Research on composition-process-property prediction of die casting Al alloys via combining feature creation and attention mechanisms

In industrial applications, the mechanical properties of die casting Al alloy parts are influenced by a combination of material composition and forming process. This study establishes a performance data-driven framework that takes into account balanced weights of alloy compositions and process parameters. First, a compositional coefficient normalization method is proposed, in which the electronegativity, thermal, and physical descriptors are dot-multiplied into the coefficient matrix to build alloy factors for coupled compositional information. Higher prediction accuracy and lower risk of overfitting validate the feasibility of the method. Subsequently, to improve the performance of machine learning (ML) model for prediction of mechanical properties, a feature engineering method embedded in attention mechanism is used. The experimental results with high accuracy confirm the ability of the ML model to reproduce the relationships between alloy properties and balance their sensitivities. Finally, contour plots of the characterized parameters are constructed for predicting mechanical property intervals and guiding parameter design. This study innovatively applies the normalization method to the field of die casting aluminum alloy feature optimization. And combined with the attention mechanism to improve the prediction accuracy between material features and target properties, which may also be extended to other alloys.

On the Use of Kullback–Leibler Divergence for Kernel Selection and Interpretation in Variational Autoencoders for Feature Creation

This study presents a novel approach for kernel selection based on Kullback–Leibler divergence in variational autoencoders using features generated by the convolutional encoder. The proposed methodology focuses on identifying the most relevant subset of latent variables to reduce the model’s parameters. Each latent variable is sampled from the distribution associated with a single kernel of the last encoder’s convolutional layer, resulting in an individual distribution for each kernel. Relevant features are selected from the sampled latent variables to perform kernel selection, which filters out uninformative features and, consequently, unnecessary kernels. Both the proposed filter method and the sequential feature selection (standard wrapper method) were examined for feature selection. Particularly, the filter method evaluates the Kullback–Leibler divergence between all kernels’ distributions and hypothesizes that similar kernels can be discarded as they do not convey relevant information. This hypothesis was confirmed through the experiments performed on four standard datasets, where it was observed that the number of kernels can be reduced without meaningfully affecting the performance. This analysis was based on the accuracy of the model when the selected kernels fed a probabilistic classifier and the feature-based similarity index to appraise the quality of the reconstructed images when the variational autoencoder only uses the selected kernels. Therefore, the proposed methodology guides the reduction of the number of parameters of the model, making it suitable for developing applications for resource-constrained devices.

Recent advances in the SISSO method and their implementation in the SISSO++ code.

Accurate and explainable artificial-intelligence (AI) models are promising tools for accelerating the discovery of new materials. Recently, symbolic regression has become an increasingly popular tool for explainable AI because it yields models that are relatively simple analytical descriptions of target properties. Due to its deterministic nature, the sure-independence screening and sparsifying operator (SISSO) method is a particularly promising approach for this application. Here, we describe the new advancements of the SISSO algorithm, as implemented into SISSO++, a C++ code with Python bindings. We introduce a new representation of the mathematical expressions found by SISSO. This is a first step toward introducing "grammar" rules into the feature creation step. Importantly, by introducing a controlled nonlinear optimization to the feature creation step, we expand the range of possible descriptors found by the methodology. Finally, we introduce refinements to the solver algorithms for both regression and classification, which drastically increase the reliability and efficiency of SISSO. For all these improvements to the basic SISSO algorithm, we not only illustrate their potential impact but also fully detail how they operate both mathematically and computationally.

FIELDimageR.Extra: Advancing user experience and computational efficiency for analysis of orthomosaic from agricultural field trials

AbstractFieldimageR.Extra is an advanced R package that enhances user experience and computational efficiency in analyzing orthomosaic images from agricultural field trials. Through the integration of modern GIS libraries like terra, stars, and sf, it surpasses traditional packages like raster and sp. The FieldimageR.Extra offers interactive visualization, vector feature creation, and spatial data editing within the R environment through the integration of mapview and mapedit packages. Notably, FieldimageR.Extra introduces new functions like fieldShape_render and fieldShape_edit, enabling flexible plot grid generation and editing capabilities. Comparative evaluations demonstrate the superior performance of FieldimageR.Extra in handling spatial data, crop and extract operations, and plot grid generation. With its comprehensive features, FieldimageR.Extra stands as a valuable addition to the FieldimageR R package, offering researchers efficient tools for unmanned aerial vehicle‐based orthomosaic image analysis in agricultural research. FIELDimageR.Extra is publicly available as a GitHub repository (https://github.com/filipematias23/FIELDimageR.Extra).

Automated Aviation Wind Nowcasting: Exploring Feature-Based Machine Learning Methods

Wind factors significantly influence air travel, and extreme conditions can cause operational disruptions. Machine learning approaches are emerging as a valuable tool for predicting wind patterns. This research, using Madeira International Airport as a case study, delves into the effectiveness of feature creation and selection for wind nowcasting, focusing on predicting wind speed, direction, and gusts. Data from four sensors provided 56 features to forecast wind conditions over intervals of 2, 10, and 20 min. Five feature selection techniques were analyzed, namely mRMR, PCA, RFECV, GA, and XGBoost. The results indicate that combining new wind features with optimized feature selection can boost prediction accuracy and computational efficiency. A strong spatial correlation was observed among sensors at different locations, suggesting that the spatial-temporal context enhances predictions. The best accuracy for wind speed forecasts yielded a mean absolute percentage error of 0.35%, 0.53%, and 0.63% for the three time intervals, respectively. Wind gust errors were 0.24%, 0.33%, and 0.38%, respectively, while wind direction predictions remained challenging with errors above 100% for all intervals.

3D Reconstruction and Measurement Analysis of a Dense Point Cloud Fused with a Depth Image

To solve the problems of holes, noise, and texture information missing in the traditional incremental reconstruction of complex surface objects, a 3D reconstruction method of depth image fusion surface dense point clouds is proposed, and texture feature creation is combined to obtain a 3D reconstruction model that takes into account the main body and details of the reconstructed object. First, the mechanism of surface dense reconstruction based on the patch-based multiview stereo (PMVS) algorithm is analyzed. Combined with the principle of view angle selection of stereo images, surface point cloud density reconstruction is performed. Then, the depth value is optimized by the region growing method, and the optimization model is established. The depth image is fused into a dense surface, and the reconstructed part is supplemented by the depth information. Finally, the Markov random field (MRF) is introduced to describe the richness of image details, and combined with the calculating method of the area coordinate, the texture coordinates are accurately calculated to reproduce the texture details of the 3D reconstruction model. 3D reconstruction experiments are performed on multiple indoor and outdoor model surfaces, and the experimental results show that the proposed method can achieve complete and accurate reconstruction of complex surface objects. Our method provides technical support for complex surface topography detection and has industrial practical significance.

Attention‐based model for dynamic IR drop prediction with multi‐view features

AbstractDynamic IR drop prediction based on machine learning has been studied in recent years. However, most proposed models used all input features extracted from circuits or manually selected parts of raw features as inputs, which failed to differentiate the order of priority among input features in a flexible manner. In this paper, QuantumForest to vector‐based dynamic IR drop prediction is introduced. With the sparse attention mechanism brought by QuantumForest, important attributes of circuits are weighed more heavily than others. A new multi‐view feature creation method is also proposed and a novel regional distance feature is built up subsequently. The performance is evaluated on two chip designs with real simulation vectors. The experiment results indicate that the prediction result of the method outperforms other prominent methods for dealing with machine learning based IR drop analysis, reaching an average MAE of only 1.457 on two designs.