Influence functions for linear discriminant analysis: Sensitivity analysis and efficient influence diagnostics

Whilst influence functions for linear discriminant analysis (LDA) have been found a single when dealing with two groups, until now these not derived in the setting of general number groups. In this paper we explore relationship between Sliced Inverse Regression (SIR) and LDA, exploit to develop LDA from those already SIR. These can be used understand robustness properties also detect influential observations practice. We illustrate usefulness via their application real data set.