Development and validation of predictive models for diabetic retinopathy using machine learning

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(2), P. e0318226 - e0318226

Published: Feb. 24, 2025

Objective This study aimed to develop and compare machine learning models for predicting diabetic retinopathy (DR) using clinical biochemical data, specifically logistic regression, random forest, XGBoost, neural networks. Methods A dataset of 3,000 patients, including 1,500 with DR, was obtained from the National Population Health Science Data Center. Significant predictors were identified, four predictive developed. Model performance assessed accuracy, precision, recall, F1-score, area under curve (AUC). Results Random forest XGBoost demonstrated superior performance, achieving accuracies 95.67% 94.67%, respectively, AUC values 0.991 0.989. Logistic regression yielded an accuracy 76.50% (AUC: 0.828), while networks achieved 82.67% 0.927). Key included 24-hour urinary microalbumin, HbA1c, serum creatinine. Conclusion The highlights as effective tools early DR detection, emphasizing importance renal glycemic markers in risk assessment. These findings support integration into decision-making improved patient outcomes diabetes management.

Language: Английский

---

Predicting pregnancy at the first year following metabolic-bariatric surgery: development and validation of machine learning models

Surgical Endoscopy, Journal Year: 2025, Volume and Issue: unknown

Published: March 10, 2025

Language: Английский

---

Machine learning analysis of financial behavior: A study of Gen Y and Gen Z preferences

Multidisciplinary Science Journal, Journal Year: 2025, Volume and Issue: 7(8), P. 2025380 - 2025380

Published: Feb. 12, 2025

To be financially sound and successful, it is necessary for an individual to have fundamental financial insight. Globalization leading several transformations across the globe. This leads both good bad impact on behavior of individuals. Even though globalization has created multiple growth avenues, also led creation a consumptive lifestyle among people. unhealthy behaviors like, overspending, lack budget saving. Current literature suggests that Generation Y Z more disposable income than any other previous generation. Thus, these two generation’s consumption investment choices certainly shape economy. Many studies in past been conducted study unique characteristics Z, current emphasizes studying Z. The aims at comparing preferences generations using machine learning algorithms. major findings reveal reveals distinctive predictive algorithms different categories Gen spending prediction, Random Forest emerges as algorithm choice while Support Vector Machine stands out with prevailing overall. These are instrumental understanding distinct patterns exhibited by providing valuable guidance institutions, planners Advisors tailoring strategies services.

Language: Английский

---

Machine Learning-Driven Metabolic Syndrome Prediction: An International Cohort Validation Study

Healthcare, Journal Year: 2024, Volume and Issue: 12(24), P. 2527 - 2527

Published: Dec. 13, 2024

Background/Objectives: This study aimed to develop and validate a machine learning (ML)-based metabolic syndrome (MetS) risk prediction model. Methods: We examined data from 6155 participants of the China Health Retirement Longitudinal Study (CHARLS) in 2011. The LASSO regression feature selection identified best MetS predictors. Nine ML-based algorithms were adopted build predictive models. model performance was validated using cohort Korea National Nutrition Examination Survey (KNHANES) (n = 5297), United Kingdom (UK) Biobank 218,781), (NHANES) 2549). Results: multilayer perceptron (MLP)-based performed CHARLS (AUC 0.8908; PRAUC 0.8073), logistic KNHANES 0.9101, 0.8116), xgboost UK 0.8556, 0.6246), MLP NHANES 0.9055, 0.8264). Conclusions: Our MLP-based has potential serve as clinical application for detecting different populations.

Language: Английский

---

Smart Home Automation with IoT-driven IR Blasters: Enhancing Control of Legacy Appliances with Predictive Usage Pattern using Machine Learning

Published: Sept. 18, 2024

Language: Английский

---