Risk assessment of customer churn in telco using FCLCNN-LSTM model

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: 248, P. 123352 - 123352

Published: Feb. 2, 2024

Language: Английский

---

A survey on sparrow search algorithms and their applications

International Journal of Systems Science, Journal Year: 2023, Volume and Issue: 55(4), P. 814 - 832

Published: Dec. 28, 2023

The sparrow search algorithm (SSA) is an efficient swarm-intelligence-based that has made some significant advances since its introduction in 2020. A detailed overview of the basic SSA and several SSA-based variants presented this paper. To be specific, first, principle introduced including mechanism implementation process. Second, many improved SSAs are reviewed hybrid, chaotic, adaptive, binary multi-objective SSAs. In addition, applications real scenarios such as machine learning areas, energy systems, path planning image processing. Finally, further research directions discussed. This survey paper aims to provide a timely review on latest developments

Language: Английский

---

Digital Mapping and Scenario Prediction of Soil Salinity in Coastal Lands Based on Multi-Source Data Combined with Machine Learning Algorithms

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(14), P. 2681 - 2681

Published: July 22, 2024

Salinization is a major soil degradation process threatening ecosystems and posing great challenge to sustainable agriculture food security worldwide. This study aimed evaluate the potential of state-of-the-art machine learning algorithms in salinity (EC1:5) mapping. Further, we predicted distribution patterns under different future scenarios Yellow River Delta. A geodatabase comprising 201 samples 19 conditioning factors (containing data based on remote sensing images such as Landsat, SPOT/VEGETATION PROBA-V, SRTMDEMUTM, Sentinel-1, Sentinel-2) was used compare predictive performance empirical bayesian kriging regression, random forest, CatBoost models. The model exhibited highest with both training testing datasets, an average MAE 1.86, RMSE 3.11, R2 0.59 datasets. Among explanatory factors, Na most important for predicting EC1:5, followed by normalized difference vegetation index organic carbon. Soil EC1:5 predictions suggested that Delta region faces severe salinization, particularly coastal zones. three increases carbon content (1, 2, 3 g/kg), 2 g/kg scenario resulted best improvement effect saline–alkali soils > ds/m. Our results provide valuable insights policymakers improve land quality plan regional agricultural development.

Language: Английский

---

A Feature Importance-Based Multi-Layer CatBoost for Student Performance Prediction

IEEE Transactions on Knowledge and Data Engineering, Journal Year: 2024, Volume and Issue: 36(11), P. 5495 - 5507

Published: May 3, 2024

Student performance prediction is vital for identifying at-risk students and providing support to help them succeed academically. In this paper, we propose a feature importance-based multi-layer CatBoost approach predict the students' grade in period exam. The idea construct structure with increasingly important features layer by layer. Specifically, importance are first calculated sorted ascending order. each layer, least accumulated until reaching given threshold. Then, these selected used training CatBoost. Next, trained utilized generate that adds set their within After that, all train next This process repeated used. results show proposed model has best performance. Moreover, statistical test conducted based on 20-runs of experiments validates significant superiority our over compared models demonstrates efficacy enhancing model. indicates can decision makers educational quality.

Language: Английский

---

A data-driven soft sensor model for coal-fired boiler SO2 concentration prediction with non-stationary characteristic

Energy, Journal Year: 2024, Volume and Issue: 300, P. 131522 - 131522

Published: May 7, 2024

Language: Английский

---

Decoding children dental health risks: a machine learning approach to identifying key influencing factors

Frontiers in Artificial Intelligence, Journal Year: 2024, Volume and Issue: 7

Published: June 17, 2024

Introduction and objectives This study investigates key factors influencing dental caries risk in children aged 7 under using machine learning techniques. By addressing caries’ prevalence, it aims to enhance early identification preventative strategies for high-risk individuals. Methods Data from clinical examinations of 356 were analyzed Logistic Regression, Decision Trees, Random Forests models. These models assessed the influence dietary habits, fluoride exposure, socio-economic status on risk, emphasizing accuracy, precision, recall, F1 score, AUC metrics. Results Poor oral hygiene, high sugary diet, low exposure identified as significant factors. The Forest model demonstrated superior performance, illustrating potential complex health data analysis. Our SHAP analysis poor Conclusion Machine effectively identifies quantifies children. approach supports targeted interventions preventive measures, improving pediatric outcomes. Clinical significance leveraging pinpoint crucial factors, this research lays groundwork data-driven strategies, potentially reducing prevalence promoting better

Language: Английский

---

Enhanced cardiovascular disease prediction through self-improved Aquila optimized feature selection in quantum neural network & LSTM model

Frontiers in Medicine, Journal Year: 2024, Volume and Issue: 11

Published: June 20, 2024

Introduction Cardiovascular disease (CVD) stands as a pervasive catalyst for illness and mortality on global scale, underscoring the imperative sophisticated prediction methodologies within ambit of healthcare data analysis. The vast volume medical available necessitates effective mining techniques to extract valuable insights decision-making prediction. While machine learning algorithms are commonly employed CVD diagnosis prediction, high dimensionality datasets poses performance challenge. Methods This research paper presents novel hybrid model predicting CVD, focusing an optimal feature set. proposed encompasses four main stages namely: preprocessing, extraction, selection (FS), classification. Initially, preprocessing eliminates missing duplicate values. Subsequently, extraction is performed address issues, utilizing measures such central tendency, qualitative variation, degree dispersion, symmetrical uncertainty. FS optimized using self-improved Aquila optimization approach. Finally, hybridized combining long short-term memory quantum neural network trained selected features. An algorithm devised optimize LSTM model’s weights. Performance evaluation approach conducted against existing models specific measures. Results Far dataset-1, accuracy-96.69%, sensitivity-96.62%, specifity-96.77%, precision-96.03%, recall-97.86%, F1-score-96.84%, MCC-96.37%, NPV-96.25%, FPR-3.2%, FNR-3.37% dataset-2, accuracy-95.54%, sensitivity-95.86%, specifity-94.51%, F1-score-96.94%, MCC-93.03%, NPV-94.66%, FPR-5.4%, FNR-4.1%. findings this study contribute improved by efficient with Discussion We have proven that our method accurately predicts cardiovascular unmatched precision conducting extensive experiments validating methodology large dataset patient demographics clinical factors. QNN frameworks tuning increase forecast accuracy reveal risk-related physiological pathways. Our shows how advanced computational tools may alter sickness management, contributing emerging field in healthcare. used revolutionary produced significant advances

Language: Английский

---

Prediction of heart failure risk factors from retinal optical imaging via explainable machine learning

Frontiers in Medicine, Journal Year: 2025, Volume and Issue: 12

Published: March 17, 2025

Over 64 million people worldwide are affected by heart failure (HF), a condition that significantly raises mortality and medical expenses. In this study, we explore the potential of retinal optical coherence tomography (OCT) features as non-invasive biomarkers for classification subtypes: left ventricular (LVHF), congestive (CHF), unspecified (UHF). By analyzing measurements from eye, right both eyes, aim to investigate relationship between ocular indicators using machine learning (ML) techniques. We conducted nine experiments compare normal individuals against LVHF, CHF, UHF patients, OCT each eye separately in combination. Our analysis revealed thickness metrics, particularly ISOS-RPE macular various regions, were reduced patients. Logistic regression, CatBoost, XGBoost models demonstrated robust performance, with notable accuracy area under curve (AUC) scores, especially classifying CHF UHF. Feature importance highlighted key parameters, such inner segment-outer segment pigment epithelium (ISOS-RPE) nuclear layer external limiting membrane (INL-ELM) thickness, crucial detection. The integration explainable artificial intelligence further enhanced model interpretability, shedding light on biological mechanisms linking changes pathology. findings suggest features, when derived have significant tools early detection failure. These insights may aid developing wearable, portable diagnostic systems, providing scalable solutions personalized healthcare, improving clinical outcomes

Language: Английский

---

Automatic Feature Selection for Imbalanced Echocardiogram Data Using Event-Based Self-Similarity

Diagnostics, Journal Year: 2025, Volume and Issue: 15(8), P. 976 - 976

Published: April 11, 2025

Background and Objective: Using echocardiogram data for cardiovascular disease (CVD) can lead to difficulties due imbalanced datasets, leading biased predictions. Machine learning models enhance prognosis accuracy, but their effectiveness is influenced by optimal feature selection robust classification techniques. This study introduces an event-based self-similarity approach automatic data. Critical features correlated with progression were identified leveraging patterns. used dataset, visual presentations of high-frequency sound wave signals, patients heart who are treated using three treatment methods: catheter ablation, ventricular defibrillator, drug control-over the course years. Methods: The dataset was classified into nine categories Recursive Feature Elimination (RFE) applied identify most relevant features, reducing model complexity while maintaining diagnostic accuracy. models, including XGBoost CATBoost, trained evaluated. Results: Both achieved comparable accuracy values, 84.3% 88.4%, respectively, under different normalization To further optimize performance, combined a voting ensemble, improving predictive Four essential features-age, aorta (AO), left (LV), atrium (LA)-were as critical found in Random Forest (RF)-voting ensemble classifier. results underscore importance techniques handling robustness, bias automated systems. Conclusions: Our findings highlight potential machine learning-driven analysis patient care providing accurate, data-driven assessments.

Language: Английский

---

Investigation of machine learning algorithms on heart disease through dominant feature detection and feature selection

Signal Image and Video Processing, Journal Year: 2024, Volume and Issue: 18(4), P. 3943 - 3955

Published: March 4, 2024

Language: Английский

---