Improving Adaptive Penalized Likelihood in Poisson Regression Model
摘要
One for difficulties in utilizing the Poisson reg. Mod. Where explanatory variable are associated is reducing the high dimensional count data using penalized Poisson reg. In order to address the drawbacks of adaptive elastic net in variable select where there is little to medium correlation between explanatory variables, an improving of adaptive elastic net (IAEN) was presented in this study to account for grouping effects in a Poisson regression model. Through simulation, the IAEN performance was showcased. In comparison to other existing penalized approaches, our simulation tests demonstrate that IAEN has an advantage small and medium and very correlate variable in terms for each prediction with variable select accuracy. Thus, IAEN is a trustworthy penalized approach for grouping effects poison reg. Mod., we can conclude.