Comparison of multivariate adaptive Poisson Regression spline and multivariate adaptive generalized Poisson Regression spline

作     者：Septia Devi Prihastuti Yasmirullah Bambang Widjanarko Otok Jerry Dwi Trijoyo Purnomo Dedy Dwi Prastyo 

作者机构：Data Science Technology Study Program Faculty of Advanced Technology and Multidiscipline Universitas Airlangga Surabaya 60115 Indonesia Department of Statistics Faculty of Science and Data Analytics Institut Teknologi Sepuluh Nopember Surabaya 60111 Indonesia 

出 版 物：《AIP Conference Proceedings》 

年 卷 期：2023年第2536卷第1期

学科分类：07[理学] 0702[理学-物理学] 

摘      要：The multivariate adaptive regression (MARS) spline is a statistical method introduced by Friedman in 1991. The MARS method has several advantages, especially for data that consider the potential nonlinear pattern. These produce a continuous model on knots where the knot selection process is performed via adaptive procedures, capturing additive effects and interactions between predictor variables, and applied to data modeling with numerical and categorical responses. However, the analyzed data is not only numerical or categorical data, but also count type data. Count type data is often found in the health sector, for example, the number of data on tuberculosis. It invites to modify the MARS model with Poisson regression and generalized Poisson regression, i.e. Multivariate Adaptive Poisson Regression Spline (MAPRS) and Multivariate Adaptive Generalized Poisson Regression Spline (MAGPRS) . In addition, this study aimed to compare MAPRS and MAGPRS when applied to tuberculosis data. By applying MAPRS and MAGPRS to model the number of tuberculosis in Lamongan regency, the MAPRS method outperforms the MAGPRS method. The MAPRS method has a lower RMSE value than the MAGPRS method.