SIMPLE ALGORITHM TO CONSTRUCT CIRCULAR CONFIDENCE REGIONS IN CORRESPONDENCE ANALYSIS USING R

Abstract

Correspondence analysis has been widely applied in various fields as a graphical method to depict the association structure between two categorical random variables on a low-dimensional plot. This study built a simple algorithm to determine the principal coordinates and construct the circular confidence regions on the correspondence plot. In this algorithm, the determination of the standard residual matrix and the principal coordinates is built directly from the contingency table (without calculating a correspondence matrix). The algorithm was developed using R and applied to data on Covid-19 cases in West Java.