Multiple Algorithms Research Articles

Depression Prediction Model based on Ensemble Learning Classifier

Objective: The main objective of this research was to develop a suitable prediction model to classify the symptoms of depression experienced by people. Methodology: This research incorporates the dataset of the “Centres for Disease Control and Prevention National Health and Nutrition Examination Survey,” which was available on GitHub. After that, pre-processing of the dataset was done using the infinite latent feature selection (ILFS) algorithm to extract the appropriate features from the dataset. After that, the dataset was split into 70:30 ratios. About 70% of the data is employed for training the machine learning algorithm, while the remaining 30% is used for testing. Further, for classification purposes, multiple machine learning algorithms were taken into consideration using the ensemble learning classifier, and the final prediction was made based on the voting classifier. The voting classifier helps determine which machine learning algorithm has superior performance over other machine learning algorithms. Finally, the proposed approach is evaluated via simulation on the dataset, with a number of metrics for performance being obtained. Finding: The result demonstrates that the suggested model achieves a high-value accuracy of 0.9166, a prediction value of 0.9177, a sensitivity value of 0.9984, a specificity value of 0.9984, and an F1-Score value of 0.9564. Finally, the comparison shows that the proposed mode outperforms machine learning algorithms, namely, ANN, XGBoost, Adaboost, random forest, stochastic gradient boosting, and SVM. Novelty: The proposed model was classified using various machine learning algorithms in place of one. Therefore, the accuracy of the proposed model was high. Keywords: Anxiety, Classification, Depression, Ensemble Learning, ILFS, Linear Regression, ML, Prediction, Stress

Hybrid wind speed optimization forecasting system based on linear and nonlinear deep neural network structure and data preprocessing fusion

Wind speed time series forecasting has been widely used in wind power generation. However, the nonlinear and non-stationary characteristics of wind speed make accurate wind speed forecasting a difficult task. In recent years, the rapid development of artificial intelligence and machine learning technology provides a new solution to the problem of wind speed forecasting. Combining the advance of artificial intelligence and data analysis strategy, this paper proposes a wind speed forecasting system based on multi-model fusion and integrated learning. Considering the differences in data observation and training principles of various algorithms and exploiting the advantages of each model, a wind speed forecasting system with multiple machine learning algorithms embedded in integrated learning is constructed, which includes three modules: data preprocessing, optimization and forecasting. The data preprocessing module can conduct quantitative analysis through input data decomposition and feature extraction, and the combination of multi-objective intelligent optimization algorithm and combined forecasting method can effectively forecast the wind speed time series. The validity of the algorithm is verified using the data of Shandong wind farm in China. The forecasting results show that compared with the traditional single model forecasting, the proposed integrated wind speed forecasting system based on the fusion of multiple models has higher forecasting accuracy.

Tolerance to asynchrony in algorithms for multiplication and modulo

In this article, we study some parallel processing algorithms for multiplication and modulo operations. We demonstrate that the state transitions that are formed under these algorithms satisfy lattice-linearity, where these algorithms induce a lattice among the global states. Lattice-linearity implies that these algorithms can be implemented in asynchronous environments, where the nodes are allowed to read old information from each other. It means that these algorithms are guaranteed to converge correctly without any synchronization overhead. These algorithms also exhibit snap-stabilizing properties, i.e., starting from an arbitrary state, the sequence of state transitions made by the system strictly follows its specification.

Tumor-infiltrating B cell-related lncRNA crosstalk reveals clinical outcomes and tumor immune microenvironment in ovarian cancer based on single-cell and bulk RNA-sequencing

BackgroundThe tumor immune microenvironment (TIME) plays a pivotal role in determining ovarian cancer (OC) prognosis. Long non-coding RNAs (lncRNAs) are key regulators of immune response and tumor progression in OC. Among these, tumor-infiltrating B cells represent an emerging target in immune response pathways. However, the specific involvement of B cell-related lncRNAs (BCRLs) in OC remains unclarified. MethodsLeveraging single-cell and bulk RNA-sequencing data, correlation analysis identified BCRLs in ovarian serous cystadenocarcinoma (OV) from the TCGA database. Subsequently, BCRLIs were filtered through COX survival analysis and the LASSO algorithm, leading to the development of a B cell-related lncRNA scoring system (BCRLss). The predictive accuracy of BCRLss for prognosis in TCGA-OV was assessed and externally validated in an independent cohort. Functional enrichment analyses were conducted to elucidate biological pathways associated with risk subgroups. Additionally, the relationship between BCRLss and TIME was investigated through multiple algorithms and consensus clustering, uncovering potential immune response targets. Drug sensitivity analyses further identified potential therapeutic options tailored to risk subgroups. The highest risk score lncRNA was selected for in vitro validation. ResultsThe BCRLss was constructed using six BCRLIs. Survival analysis revealed an improved prognosis in the low-risk group, with results corroborated by external validation in the ICGC-OV cohort. ROC analysis and nomogram construction confirmed the strong prognostic accuracy of BCRLss. Enrichment analysis highlighted associations between risk subgroups and tumor immune pathways, with the low-risk group demonstrating a more robust immune response and elevated expression of immune checkpoint-related genes. Drug sensitivity tests revealed notable differences across risk subgroups. In vitro experiments confirmed elevated LINC01150 expression in OC cells, and LINC01150 knockdown significantly inhibited the proliferation, invasion, and migration of SKOV3 cells. ConclusionsIn conclusion, BCRLss provides a reliable prognostic tool for predicting clinical outcomes and the immune landscape of patients with OC, offering valuable guidance for immunotherapy target selection and personalized treatment strategies.

Integrative analysis of single-cell and bulk multi-omics data to reveal subtype-specific characteristics and therapeutic strategies in clear cell renal cell carcinoma patients.

Background: Kidney renal clear cell carcinoma (KIRC) is the most prevalent subtype of malignant renal cell carcinoma and is well known as a common genitourinary cancer. Stratifying tumors based on heterogeneity is essential for better treatment options. Methods: In this study, consensus clusters were constructed based on gene expression, DNA methylation, and gene mutation data, which were combined with multiple clustering algorithms. After identifying two heterogeneous subtypes, we analyzed the molecular characteristics, immunotherapy response, and drug sensitivity differences of each subtype. And we further integrated bulk data and single-cell RNA sequencing (scRNA-Seq) data to infer the immune cell composition and malignant tumor cell proportion of subtype-related cell subpopulations. Results: Among the two identified consensus subtypes (CS1 and CS2), CS1 was enriched in more inflammation-related and oncogenic pathways than CS2. Simultaneously, CS1 showed a worse prognosis and we found more copy number variations and BAP1 mutations in CS1. Although CS1 had a high immune infiltration score, it exhibited high expression of suppressive immune features. Based on the prediction of immunotherapy and drug sensitivity, we inferred that CS1 may respond poorly to immunotherapy and be less sensitive to targeted drugs. The analysis of bulk data integrated with single-cell data further reflected the high expression of inhibitory immune features in CS1 and the high proportion of malignant tumor cells. And CS2 contained a large number of plasmacytoid B cells, presenting an activated immune microenvironment. Finally, the robustness of our subtypes was successfully validated in four external datasets. Conclusion: In summary, we conducted a comprehensive analysis of multi-omics data with 10 clustering algorithms to reveal the molecular characteristics of KIRC patients and validated the relevant conclusions by single-cell analysis and external data. Our findings discovered new KIRC subtypes and may further guide personalized and precision treatments.

Objective quantification of sound sensory attributes in side-by-side vehicles using multiple linear regression models

The evaluation of sound quality is a pivotal area of research within audio and acoustics. The sound quality evaluation methods commonly used include both objective and subjective, the latter being time-consuming and costly as they rely on listening tests. This research work aims to investigate the use of predictive sound quality models as a way to objectively assess the Desire-to-buy of side-by-side vehicles, in a more efficient, faster, and less costly way than conventional methods. Multiple linear regression algorithms were used to validate the objective models derived from objective physical metrics and perceptual psycho-physical metrics. The sensory profile objective models reported in this paper were constructed using parsimonious linear Lasso and Elastic-net algorithms. Our results show that linear objective models effectively account for each of the perceptual attributes of the sensory profiles and the Desire-to-buy, while only requiring a few physical and psychophysical metrics.

Spatial heterogeneity of ecosystem service bundles and the driving factors in the Beijing-Tianjin-Hebei region

Defining a multi-service spatial pattern and ecosystem packages is the key to ensuring regional ecological security and sustainable development. Ecosystem service bundles can effectively discern the interactions among multiple services and regionalize ecological functional zones. Aiming at a range of environmental and ecological challenges in the Beijing-Tianjin-Hebei region, such as water scarcity, soil erosion, land degradation, poor forest and grassland quality, and shrinking agricultural land, the spatial pattern and driving forces of eight crux ecosystem services, including product supply, soil conservation, water conservation, carbon storage, sand fixation, environment purification, biological diversity and recreation culture, were analyzed by combining multiple model algorithms. Additionally, spatial system clustering was employed to identify ecosystem services bundles and optimize the ecological functional zoning. The results revealed significant spatial heterogeneity of the eight ecosystem services in the Beijing-Tianjin-Hebei region. High product supply concentrated in plain regions, while the rest seven ecosystem services were dominant in in mountainous areas. The region can be classified into four service bundles: ecological conservation service (ECS), ecological restoration service (ERS), product supply service (PSS) and ecological fragile service (EFS). ECS and ERS are dominated by regulating and supporting services, which are primarily driven by climatic factors and NDVI. ERS exhibits relatively low service level and is the crucial area for future ecological restoration. PSS is dominated by food supply, with land use change being the primary driving force. At present, with the rapid urbanization, various ecosystem services provided by EFS is inadequate. In the future, urban ecological construction should be strengthened to enhance regional ecosystem services capabilities.

Study on Predictive Modeling for Loan Approval

In the financial business, authorizing a loan is a critical decision-making step that often necessitates a human review of applicants' financial information and credit history. However, as massive datasets become more readily available and computing technology advances, machine learning approaches will become more powerful tools for improving the precision and effectiveness of loan approval processes. This study investigates the use of machine learning techniques combined with predictive modeling to automate and optimize loan approval processes. Multiple algorithms analyze crucial features such as job status, credit score, income, and loan amount to determine if a loan will be accepted or rejected. These approaches include logistic regression, decision trees, random forests, and neural networks. The goal of the research is to create an automated loan approval system.

Fatty acid metabolism prognostic signature predicts tumor immune microenvironment and immunotherapy, and identifies tumorigenic role of MOGAT2 in lung adenocarcinoma.

Aberrant fatty acid metabolism (FAM) plays a critical role in the tumorigenesis of human malignancies. However, studies on its impact in lung adenocarcinoma (LUAD) are limited. We developed a prognostic signature comprising 10 FAM-related genes (GPR115, SOAT2, CDH17, MOGAT2, COL11A1, TCN1, LGR5, SLC34A2, RHOV, and DKK1) using data from LUAD patients in The Cancer Genome Atlas (TCGA). This signature was validated using six independent LUAD datasets from the Gene Expression Omnibus (GEO). Patients were classified into high- and low-risk groups, and overall survival (OS) was compared by Kaplan-Meier analysis. The signature's independence as a prognostic indicator was assessed after adjusting for clinicopathological features. Receiver operating characteristic (ROC) analysis validated the signature. Tumor immune microenvironment (TIME) was analyzed using ESTIMATE and multiple deconvolution algorithms. Functional assays, including CCK8, cell cycle, apoptosis, transwell, and wound healing assays, were performed on MOGAT2-silenced H1299 cells using CRISPR/Cas9 technology. Low-risk group patients exhibited decreased OS. The signature was an independent prognostic indicator and demonstrated strong risk-stratification utility for disease relapse/progression. ROC analysis confirmed the signature's validity across validation sets. TIME analysis revealed higher infiltration of CD8+ T cells, natural killers, and B cells, and lower tumor purity, stemness index, and tumor mutation burden (TMB) in low-risk patients. These patients also showed elevated T cell receptor richness and diversity, along with reduced immune cell senescence. High-risk patients exhibited enrichment in pathways related to resistance to immune checkpoint blockades, such as DNA repair, hypoxia, epithelial-mesenchymal transition, and the G2M checkpoint. LUAD patients receiving anti-PD-1 treatment had lower risk scores among responders compared to non-responders. MOGAT2 was expressed at higher levels in low-risk LUAD patients. Functional assays revealed that MOGAT2 knockdown in H1299 cells promoted proliferation and migration, induced G2 cell cycle arrest, and decreased apoptosis. This FAM-related gene signature provides a valuable tool for prognostic stratification and monitoring of TIME and immunotherapy responses in LUAD. MOGAT2 is identified as a potential anti-tumor regulator, offering new insights into its role in LUAD pathogenesis.

Machine learning-derived peripheral blood transcriptomic biomarkers for early lung cancer diagnosis: Unveiling tumor-immune interaction mechanisms.

Lung cancer continues to be the leading cause of cancer-related mortality worldwide. Early detection and a comprehensive understanding of tumor-immune interactions are crucial for improving patient outcomes. This study aimed to develop a novel biomarker panel utilizing peripheral blood transcriptomics and machine learning algorithms for early lung cancer diagnosis, while simultaneously providing insights into tumor-immune crosstalk mechanisms. Leveraging a training cohort (GSE135304), we employed multiple machine learning algorithms to formulate a Lung Cancer Diagnostic Score (LCDS) based on peripheral blood transcriptomic features. The LCDS model's performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC) in multiple validation cohorts (GSE42834, GSE157086, and an in-house dataset). Peripheral blood samples were obtained from 20 lung cancer patients and 10 healthy control subjects, representing an in-house cohort recruited at the Sixth People's Hospital of Chengdu. We employed advanced bioinformatics techniques to explore tumor-immune interactions through comprehensive immune infiltration and pathway enrichment analyses. Initial screening identified 844 differentially expressed genes, which were subsequently refined to 87 genes using the Boruta feature selection algorithm. The random forest (RF) algorithm demonstrated the highest accuracy in constructing the LCDS model, yielding a mean AUC of 0.938. Lower LCDS values were significantly associated with elevated immune scores and increased CD4+ and CD8+ T-cell infiltration, indicative of enhanced antitumor-immune responses. Higher LCDS scores correlated with activation of hypoxia, peroxisome proliferator-activated receptor (PPAR), and Toll-like receptor (TLR) signaling pathways, as well as reduced DNA damage repair pathway scores. Our study presents a novel, machine learning-derived peripheral blood transcriptomic biomarker panel with potential applications in early lung cancer diagnosis. The LCDS model not only demonstrates high accuracy in distinguishing lung cancer patients from healthy individuals but also offers valuable insights into tumor-immune interactions and underlying cancer biology. This approach may facilitate early lung cancer detection and contribute to a deeper understanding of the molecular and cellular mechanisms underlying tumor-immune crosstalk. Furthermore, our findings on the relationship between LCDS and immune infiltration patterns may have implications for future research on therapeutic strategies targeting the immune system in lung cancer.

Towards the construction of regional marine radiocarbon calibration curves: an unsupervised machine learning approach

Abstract. Radiocarbon may serve as a powerful dating tool in palaeoceanography, but its accuracy is limited by the need to calibrate radiocarbon dates to calendar ages. A key problem is that marine radiocarbon dates must be corrected for past offsets from either the contemporary atmosphere (i.e. “reservoir age” offsets) or a modelled estimate of the global average surface ocean (i.e. delta-R offsets). This presents a challenge because the spatial distribution of reservoir ages and delta-R offsets can vary significantly, particularly over periods of major marine hydrographic and/or carbon cycle change such as the last deglaciation. Modern reservoir age and delta-R estimates therefore have limited applicability. While forward modelling of past R-age variability has been proposed as a means of resolving this problem, this requires accurate a priori knowledge of past global radiocarbon budget closure (i.e. production, and cycling), which we currently lack. In this context, the construction of empirical regional marine calibration curves could provide a way forward. However, the spatial reach of such calibrations and their robustness subject to (uncertain) temporal changes in climate and ocean circulation would need to be tested. Here, we use unsupervised machine learning techniques to define distinct regions of the surface ocean that exhibit coherent behaviour in terms of their radiocarbon age offsets from the contemporary atmosphere (R ages), regardless of the causes of R-age variability. We apply multiple algorithms (k-means, k-medoids, and hierarchical clustering) to outputs from two different numerical models spanning a range of climate states, forcings, and timescales of adjustment. Comparisons between the cluster assignments across model runs confirm some robust regional patterns that likely stem from constraints imposed by large-scale ocean and atmospheric physics. At the coarsest scale, regions of coherent R-age variability correspond to the major ocean basins. By further dividing basin-scale shape-based clusters into amplitude-based subclusters, we recover regional associations, such as increased high-latitude R ages, or the propagation of R-age anomalies from regions of deep mixing in the Southern Ocean to upwelling sites in the eastern equatorial Pacific, which cohere with known modern oceanographic processes. We show that the medoids (i.e. the most representative locations) for these regional sub-clusters provide significantly better approximations of simulated local R-age variability than constant offsets from the global surface average. This remains true when cluster assignments obtained from one model simulation are applied to simulated R-age time series from another. Further, model-based clusters are found to be broadly consistent with existing reservoir age reconstructions that span the last ∼30 kyr. We therefore propose that machine learning provides a promising approach to the problem of defining regions for which empirical marine radiocarbon calibration curves may eventually be generated.

CTPAD: an interactive web application for comprehensive transcriptomic profiling in allergic diseases

BackgroundAllergic diseases are systemic chronic inflammatory diseases associated with multiorgan damage and complex pathogenesis. Several studies have revealed the association of gene expression abnormalities with the development of allergic diseases, but the biomedical field still lacks a public platform for comprehensive analysis and visualization of transcriptomic data of allergic diseases.ObjectiveThe aim of the study is to provide a comprehensive web tool for multiple analysis in allergic diseases.MethodsWe retrieved and downloaded human and mouse gene expression profile data associated with allergic diseases from the Gene Expression Omnibus (GEO) database and standardized the data uniformly. We used gene sets obtained from the MSigDB database for pathway enrichment analysis and multiple immune infiltration algorithms for the estimation of immune cell proportion. The basic construction of the web pages was based on the Shiny framework. Additionally, more convenient features were added to the server to improve the efficiency of the web pages, such as jQuery plugins and a comment box to collect user feedback.ResultsWe developed CTPAD, an interactive R Shiny application that integrates public databases and multiple algorithms to explore allergic disease-related datasets and implement rich transcriptomic visualization capabilities, including gene expression analysis, pathway enrichment analysis, immune infiltration analysis, correlation analysis, and single-cell RNA sequencing analysis. All functional modules offer customization options and can be downloaded in PDF format with high-resolution images.ConclusionsCTPAD largely facilitates the work of researchers without bioinformatics background to enable them to better explore the transcriptomic features associated with allergic diseases. CTPAD is available at https://smuonco.shinyapps.io/CTPAD/.

Evaluation of the Multiple Regression Analysis Algorithm on Stock Market Prediction

Financial time series is one of the most challenging applications of modern time series forecasting. The financial time series is closely related to noise, non-stationary, and deterministic chaos. The characteristics suggest that no complete information can be obtained from the past behavior of financial markets to fully capture the dependency between future prices and that of the past. The data collection method was collected from the Stock Market Online Application "MetaTrader version 4" type "Daily" with a time range from "03/09/2001 to 25/07/2012", as many as 2052 data", with the attributes "Date, Open, High, Low, Close, Volume" with the main attribute "Close" using the Support vector machine algorithm, artificial neural network, and multiple linear regression. The conclusion of the value that is close to the series value is the value by testing on the support vector machine algorithm, with the parameter for the RMSE value that is close to the "0" value obtained from the measurement results on the SVM algorithm on the RBF kernel (radial base function) with a value of "gamma" γ = 100 with the value of RMSE = 0.000, and SE = 0.000. with prediction accuracy error = 0.976

Reservation and allocation model considering the user cost and the utilization of parking space

In order to alleviate the problem of imbalance between parking supply and demand by improving the utilization rate of parking resources, and considering the game relationship between user benefits and system benefits in parking allocation, the optimization objective function of maximum parking space utilization and minimum user cost in parking reservation mode is established from the two aspects which are system optimization and user optimization respectively. Then, the optimal allocation parking integer programming model (OPA) considering users' preferences is established. An Augmented Lagrangian-Alternating Direction Method of Multipliers Algorithm is designed to solve the optimal solution of the model. Finally, in order to test the validity of the model, the proposed model and the classical distribution model are compared and analyzed under different supply and demand conditions. The results show that the performance of the proposed model in three performance metrics which are parking utilization, average user cost and request acceptance rate is significantly better than that of the classical allocation model. The research results of this paper can provide theoretical reference for the management of parking reservation platform.

A Generalized Number-Theoretic Transform for Efficient Multiplication in Lattice Cryptography

The Number Theoretic Transform (NTT) has emerged as a powerful tool for efficiently computing convolutions of digital signals, due to its inherent advantages such as numerical stability, reliance on simple integer operations, and proven efficiency. Its applications have extended to accelerating polynomial multiplication in lattice-based cryptography. However, existing NTT multiplication algorithms impose restrictions on the underlying moduli, potentially affecting key and ciphertext sizes as well as computational overhead. Therefore, enabling NTT with small moduli holds significant potential for enhancing the overall system performance. This study introduces a novel reduction framework for NTT computation in cyclotomic rings employing field extensions. Our approach replaces the underlying polynomial ring with a two-dimensional isomorphic ring, effectively relaxing the restrictions imposed on the NTT moduli. The proposed framework is evaluated through two case studies relevant to the LAC and NTTRU lattice-based cryptographic schemes. Comprehensive theoretical analysis is provided, demonstrating the effectiveness of our approach in enabling NTT with small moduli and its potential to improve the efficiency of lattice-based cryptography.

Research on the Coupling and Coordination of Land Ecological Security and High-Quality Agricultural Development in the Han River Basin

This study aims to investigate the coupling and harmonization between land ecological security (LES) and high-quality agricultural development (HAD) in the Han River Basin (HRB), China, with the objective of promoting harmonious coexistence between agriculture and ecosystems. Using 17 cities in the HRB as the research objects, an evaluation index system of two systems, LES and HAD, was constructed, analyzed, and evaluated via projective tracer modeling for multiple intelligent genetic algorithms (MIGA-PTM). The degree of coupling coordination (DCC) was used to quantitatively evaluate the coupling coordination development status of the two systems, the obstacle model (OM) was used to identify the main influencing factors, and the gray predictive model first-order univariate model (GM (1, 1)) was used to predict the DCC of the LES and HAD from 2025 to 2040. The results show the following: (1) the LES and HAD levels of the 17 cities in the HRB tended to increase during the study period, and there was a large gap between cities; (2) the spatial distributions of the DCCs of the LES and HAD in the HRB were uneven, with high values in the southern and low values in the central and northern parts, and the overall degree of coupling tended to fluctuate. The overall DCC showed a fluctuating upward trend; (3) the degree of obstacles, per capita water resources, greening coverage, and rate of return on financial expenditure are the main influencing factors; and (4) the prediction results of GM (1, 1) indicate that the LES and HAD of the HRB will be close to reaching the intermediate stage of coupling in 2035. This research offers critical insights into sustainable development practices that facilitate the alignment of agricultural growth with ecological preservation.

Data-augmented machine learning scoring functions for virtual screening of YTHDF1 m6A reader protein

Machine learning is rapidly advancing the drug discovery process, significantly enhancing speed and efficiency. Innovation in computer-aided drug design is primarily driven by structure- and ligand-based approaches. When the number of known inhibitors for a target is limited, data augmentation strategies are often preferred to enhance model performance. In this study, we developed predictive machine learning models for structure-based drug discovery leveraging multiple traditional machine learning algorithms trained with target and ligand dynamics-aware datasets.To illustrate our approach, we present a composite model that combines classification and regression to predict YTHDF1 inhibitors, utilizing PLEC features. YTHDF1, a key m6A reader protein involved in mRNA translation, is implicated in various cancers, making it a promising therapeutic target. Traditional structure-based virtual screening (SBVS) using generic scoring functions has struggled to identify potent YTHDF1 inhibitors due to the protein's unique binding characteristics. To overcome this, we developed YTHDF1-specific machine learning scoring functions (MLSFs) to enhance SBVS efficacy.We employed various data augmentation techniques to generate a comprehensive dataset, incorporating multiple conformations of ligands and the YTHDF1 protein. We have trained 64 YTHDF1-specific MLSFs using four machine learning algorithms and evaluated them on ten test sets, focusing on their predictive and ranking power. Our results demonstrate that the artificial neural network with protein-ligand extended connectivity fingerprints (ANN-PLEC) outperforms other MLSFs, consistently achieving high area under the precision-recall curve (PR-AUC) of 0.87. This method shows promise for targets with limited quantities of active molecules, providing a viable path forward for drug discovery research. The ANN-PLEC scoring function is made freely available on GitHub for other researchers to access and utilize https://github.com/JuniML/SBVS-YTHDF1/.

Deep Learning-Based Fatigue Strength Prediction for Ferrous Alloy

As industrial development drives the increasing demand for steel, accurate estimation of the material’s fatigue strength has become crucial. Fatigue strength, a critical mechanical property of steel, is a primary factor in component failure within engineering applications. Traditional fatigue testing is both costly and time-consuming, and fatigue failure can lead to severe consequences. Therefore, the need to develop faster and more efficient methods for predicting fatigue strength is evident. In this paper, a fatigue strength dataset was established, incorporating data on material element composition, physical properties, and mechanical performance parameters that influence fatigue strength. A machine learning regression model was then applied to facilitate rapid and efficient fatigue strength prediction of ferrous alloys. Twenty characteristic parameters, selected for their practical relevance in engineering applications, were used as input variables, with fatigue strength as the output. Multiple algorithms were trained on the dataset, and a deep learning regression model was employed for the prediction of fatigue strength. The performance of the models was evaluated using metrics such as MAE, RMSE, R2, and MAPE. The results demonstrated the superiority of the proposed models and the effectiveness of the applied methodologies.

Early prediction of sudden cardiac death using multimodal fusion of ECG Features extracted from Hilbert–Huang and wavelet transforms with explainable vision transformer and CNN models

Background and Objective:Sudden cardiac death (SCD) is a critical health issue characterized by the sudden failure of heart function, often caused by ventricular fibrillation (VF). Early prediction of SCD is crucial to enable timely interventions. However, current methods predict SCD only a few minutes before its onset, limiting intervention time. This study aims to develop a deep learning-based model for the early prediction of SCD using electrocardiography (ECG) signals. Methods:A multimodal explainable deep learning-based model is developed to analyze ECG signals at discrete intervals ranging from 5 to 30 min before SCD onset. The raw ECG signals, 2D scalograms generated through wavelet transform and 2D Hilbert spectrum generated through Hilbert–Huang transform (HHT) of ECG signals were applied to multiple deep learning algorithms. For raw ECG, a combination of 1D-convolutional neural networks (1D-CNN) and long short-term memory networks were employed for feature extraction and temporal pattern recognition. Besides, to extract and analyze features from scalograms and Hilbert spectra, Vision Transformer (ViT) and 2D-CNN have been used. Results:The developed model achieved high performance, with accuracy, precision, recall and F1-score of 98.81%, 98.83%, 98.81%, and 98.81% respectively to predict SCD onset 30 min in advance. Further, the proposed model can accurately classify SCD patients and normal controls with 100% accuracy. Thus, the proposed method outperforms the existing state-of-the-art methods. Conclusions:The developed model is capable of capturing diverse patterns on ECG signals recorded at multiple discrete time intervals (at 5-minute increments from 5 min to 30 min) prior to SCD onset that could discriminate for SCD. The proposed model significantly improves early SCD prediction, providing a valuable tool for continuous ECG monitoring in high-risk patients.

Identifying Key Indicators for Successful Foreign Direct Investment through Asymmetric Optimization Using Machine Learning

The advancement of technology has led humanity into the era of the information society, where information drives progress and knowledge is the most valuable resource. This era involves vast amounts of data, from which stored knowledge should be effectively extracted for use. In this context, machine learning is a growing trend used to address various challenges across different fields of human activity. This paper proposes an ensemble model that leverages multiple machine learning algorithms to determine the key factors for successful foreign direct investment, which simultaneously enables the prediction of this process using data from the World Bank, covering 60 countries. This innovative model, which adds to scientific and research knowledge, employs two sets of methods—binary regression and feature selection—combined in a stacking ensemble using a classification algorithm as the combiner to enable asymmetric optimization. The proposed predictive ensemble model has been tested in a case study using a dataset compiled from World Bank data across countries worldwide. The model demonstrates better performance than each of the individual algorithms integrated into it, which are considered state-of-the-art in these methodologies. Additionally, the findings highlight three key factors for foreign direct investment from the dataset, leading to the development of an optimized prediction formula.