Health Status Prediction Supported by Explainable Machine Learning: A Voting Regression Analysis with SHAP and LIME
4th Cognitive Models and Artificial Intelligence Conference, AICCONF 2026, Prague, Czech Republic, 24 - 25 April 2026, (Full Text)
- Publication Type: Conference Paper / Full Text
- Doi Number: 10.1109/aicconf69182.2026.11600686
- City: Prague
- Country: Czech Republic
- Keywords: explainable artificial intelligence (XAI), gradient boosting regression, health status prediction, kernel ridge regression, LIME, machine learning, SHAP, support vector regression, voting regression
- Hacettepe University Affiliated: Yes
Abstract
In the scope of recent advancements in the healthcare sector, various health based analyses can be performed for individuals based on different factors. Within the scope of these analyses, accurately determining an individual's health status holds critical importance. In this study, a regression problem aimed at predicting individuals' health status is addressed and a Voting Regression model is proposed as the solution. Additionally, this model is composed of an ensemble of three different individual regression models, which are Kernel Ridge Regression, Gradient Boosting Regression, and Support Vector Regression. In the study, both the individual models and the voting model are compared quantitatively and visually in terms of performance. Furthermore, for enhancing the transparency of the voting model's decision making process, feature contribution analysis is conducted by utilizing explainable artificial intelligence (XAI) techniques, SHAP and LIME.