SOFTWARE TRICKS AND TIPS
An R function for imputation of missing cells in two-way data sets by EM-AMMI algorithm

Jakub Paderewski

Commun. Biometry Crop Sci. (2013) 8 (2), 60-69.
 

ABSTRACT
Various statistical methods for two-way classification data sets (including AMMI or GGE analyses, used in crop science for interpreting genotype-by-environment interaction) require the data to be complete, that is, not to have missing cells. If there are such, however, one might impute the missing cells. The paper offers R code for imputing missing values by the EM-AMMI algorithm. In addition, a function to check the repeatability of this algorithm is proposed. This function could be used to evaluate if the missing data were imputed reliably (unambiguously), which is important especially for small data sets

Key Words: agricultural trials; EM-AMMI; missing cells; multivariate statistics; Singular Value Decomposition; software.
 

[Full Text - pdf]