A Simulation Study Comparing Tree-Based Methods in Identifying Interactions of Continuous and Binary Variables for Prediction of Increased Risk of Disease

• DOI: 10.37256/bsr.1120232148
• Publication Date: 2023-07-07T08:41:45Z
• Authors: Sybil Prince Nelson
• Keywords: Categorical variable; Continuous variable; Tree (set theory)

Tree-based methods are commonly used to create models that predict an output based on several input variables. Classification and Regression Trees (CARTs) is a popular algorithm builds tree-like graphs for predicting continuous categorical dependent variables, but it has been shown be biased toward the inclusion of Conditional inference technique alleviate this bias. C.Logic alternative tree-based method uses Boolean logic classification trees. Previous research superior CART in identifying interactions lead increased risk disease. No comparison made between package with conditional as found called Party. In paper, simulation study compare capability these two algorithms identify binary It while both succeed correct interactions, more effective. The does better job alleviating bias variables when attempting interacting