Explainable Machine Learning for Predicting Dengue Recovery Duration: Insights from Multi-Center Clinical Data

Area of Science:

• Medical Informatics
• Public Health
• Machine Learning

Background:

• Dengue fever poses a significant public health challenge in endemic areas.
• Patient recovery duration from dengue fever is highly variable.
• Existing machine learning models for dengue prediction often lack interpretability.

Purpose of the Study:

• To develop an Explainable Artificial Intelligence (XAI) framework for analyzing dengue recovery duration.
• To identify key clinical, demographic, and contextual factors influencing dengue recovery.
• To enhance the interpretability of machine learning models in dengue research.

Main Methods:

• Utilized a multi-center clinical dataset of 100 laboratory-confirmed dengue patients.
• Developed and evaluated four machine learning models: Linear Regression, Decision Tree, Random Forest, and Neural Network.
• Employed XAI techniques (PDP, ICE, LIME) to interpret model predictions.

Main Results:

• Random Forest model achieved the best predictive performance (RMSE: 11.29 days, MAE: 9.09 days).
• Age and platelet count were consistently identified as the most influential predictors of recovery duration.
• Explainability analyses revealed significant patient-level heterogeneity in recovery patterns.

Conclusions:

• XAI models can effectively identify key predictors of dengue recovery duration.
• Older age and lower platelet counts are associated with longer recovery times.
• Dengue recovery is influenced by a complex interplay of factors, necessitating personalized healthcare approaches.