ANOVA Assisted Variable Selection in High-dimensional Multicategory Response Data

Multinomial logistic regression is preferred in the classification of multicategory response data for its ease interpretation and ability to identify associated input variables each category. However, identifying important high-dimensional poses several challenges as majority are unnecessary discriminating categories. Frequently used techniques include regularisation such Least Absolute Selection Shrinkage Operator (LASSO) sure independent screening (SIS) or combinations both. In this paper, we propose use ANOVA, assist SIS variable when multicategorical. The new approach straightforward computationally effective. Simulated without with correlation generated numerical studies illustrate methodology, results applying methods on real presented. conclusion, ANOVA performance comparable selection uncorrelated performs better combination both correlated variables.