Multivariate Outliers Detection on GGE Biplot

Abstract

Outliers are a very interesting problem, statisticians are well aware of the potential effects of outliers on data analysis, especially in multivariate data. As part of our research in modeling the two-ways table data with the Row Column Interaction Model (RCIM), we had discussed the applicable RCIM model in the Genotypes × Environments Interaction (GEI) analysis which is frequently used the Genotype and Genotype × Environments Interaction (GGE) Biplot for displaying interaction in low-dimensional space. Previously, we have studied the influenced of the outlying observations on the visualization of the interaction effects in the GGE and Genotype × Environments (GE) by RCIM modeling. Now we focused on how to detect the presence of any outliers in data of two ways table and make some suggestions for practitioners by conducted simple scheme outlying observation scenario. We also proposed the use of Robust Biplot GGE as graphical techniques for detecting outliers in our data by visualizing them on two-dimensional space.