Machine learning models show promise for TB screening using cross-cohort clinical data

Researchers have evaluated machine learning models for tuberculosis (TB) screening using clinical and demographic data from Uganda and South Africa. The study applied logistic regression, multilayer perceptrons, and convolutional neural networks to the CAGE-TB dataset, incorporating feature selection to improve model performance. While feature selection enhanced the area under the receiver operating characteristic (AUROC) curve by 2-7% for all models, the logistic regression model demonstrated consistent performance on held-out data, achieving an AUROC of 0.84 in South Africa and 0.8 in Uganda. Deeper neural networks showed less consistent results on unseen data, though the development of such classifiers for TB screening is considered viable. AI

IMPACT This research demonstrates the potential of machine learning models to improve tuberculosis screening efficiency using readily available clinical data.

RANK_REASON Academic paper detailing research methodology and results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• CAGE-TB
• convolutional neural network
• logistic regression model
• multilayer perceptron
• South Africa
• Uganda

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →