On the Selection of Regression Model Using Machine Learning

• Activity Number: 256 - Contributed Poster Presentations: Section on Statistical Learning and Data Science
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Asanao Shimokawa* and Etsuo Miyaoka
• Companies: Tokyo University of Science and Tokyo University of Science
• Keywords: Genetic algorithm; Logistic regression; Model selection

Abstract

In this study, we focus on the model selection problem in the logistic regression model. When construct the model from a data set, we need to decide which covariates should be included in the model. In addition to this, it is also need to consider interactions between covariates and their nonlinear transformations. In order to perform these operations automatically, we examine the automatic model selection method using a genetic algorithm. We propose the algorithm for this purpose and verify its performance through simulation studies.