Supervised machine learning algorithms to predict the duration and risk of long-term hospitalization in HIV-infected individuals: a retrospective study

Frontiers in Public Health, Journal Year: 2024, Volume and Issue: 11

Published: Jan. 5, 2024

Objective The study aimed to use supervised machine learning models predict the length and risk of prolonged hospitalization in PLWHs help physicians timely clinical intervention avoid waste health resources. Methods Regression were established based on RF, KNN, SVM, XGB hospital stay using RMSE, MAE, MAPE, R 2 , while classification NN, accuracy, PPV, NPV, specificity, sensitivity, kappa, visualization evaluation AUROC, AUPRC, calibration curves decision all used for internally validation. Results In regression models, model performed best internal validation (RMSE = 16.81, MAE 10.39, MAPE 0.98, 0.47) stay, NN presented good fitting stable features testing sets, with excellent accuracy (0.7623), PPV (0.7853), NPV (0.7092), sensitivity (0.8754), specificity (0.5882), kappa (0.4672), further indicated that largest AUROC (0.9779), AUPRC (0.773) well-performed curve Conclusion This showed was effective predicting PLWH. Based predictive an intelligent medical prediction system may be developed effectively HIV patients according their records, which helped reduce healthcare

Language: Английский

Systemic lupus in the era of machine learning medicine

Lupus Science & Medicine, Journal Year: 2024, Volume and Issue: 11(1), P. e001140 - e001140

Published: March 1, 2024

Artificial intelligence and machine learning applications are emerging as transformative technologies in medicine. With greater access to a diverse range of big datasets, researchers turning these powerful techniques for data analysis. Machine can reveal patterns interactions between variables large complex datasets more accurately efficiently than traditional statistical methods. approaches open new possibilities studying SLE, multifactorial, highly heterogeneous disease. Here, we discuss how methods rapidly being integrated into the field SLE research. Recent reports have focused on building prediction models and/or identifying novel biomarkers using both supervised unsupervised understanding disease pathogenesis, early diagnosis prognosis In this review, will provide an overview current gaps, challenges opportunities studies. External validation most is still needed before clinical adoption. Utilisation deep models, alternative sources health increased awareness ethics, governance regulations surrounding use artificial medicine help propel exciting forward.

Language: Английский

Global and Local Interpretable Machine Learning Allow Early Prediction of Unscheduled Hospital Readmission

Machine Learning and Knowledge Extraction, Journal Year: 2024, Volume and Issue: 6(3), P. 1653 - 1666

Published: July 17, 2024

Nowadays, most of the health expenditure is due to chronic patients who are readmitted several times for their pathologies. Personalized prevention strategies could be developed improve management these patients. The aim present work was develop local predictive models using interpretable machine learning techniques early identify individual unscheduled hospital readmissions. To do this, a retrospective, case-control study, based on information regarding patient readmission in 2018–2019, conducted. After curation initial dataset (n = 76,210), final number participants n 29,026. A analysis performed following algorithms readmissions as dependent variable. Local model-agnostic interpretability methods were also performed. We observed 13% rate cases. There statistically significant differences age and days stay (p < 0.001 both cases). logistic regression model revealed therapy (odds ratio: 3.75), diabetes mellitus history 1.14), 1.02) relevant factors. Machine yielded better results sensitivity other metrics. Following, this procedure, important factors predict Interestingly, variables like allergies adverse drug reaction antecedents relevant. Individualized prediction high sensitivity. In conclusion, our study identified influencing readmissions, emphasizing impact length stay. introduced personalized risk predicting with notable accuracy. Future research should include more clinical refine further.

Language: Английский