Articles
AN IMPROVED IMPUTATION METHOD FOR INCOMPLETE GxE TRIAL DATA FOR ASPARAGUS
Article number
589_13
Pages
111 – 116
Language
English
Abstract
Improving the analysis of incomplete genotype x environment (GxE) information has the potential to improve the efficiency of breeding programmes.
This is particularly relevant in large-scale international cultivar trials.
A two-stage imputation method is proposed which uses results from a cluster analysis; it provides more accurate estimates of missing values under random removal of data from five complete data sets in the literature.
These imputed values are superior to those obtained using other clustering based methods.
The two stage method is also considered to be superior to model based imputation as it does not rely on model selection or model fitting constraints.
This is particularly relevant in large-scale international cultivar trials.
A two-stage imputation method is proposed which uses results from a cluster analysis; it provides more accurate estimates of missing values under random removal of data from five complete data sets in the literature.
These imputed values are superior to those obtained using other clustering based methods.
The two stage method is also considered to be superior to model based imputation as it does not rely on model selection or model fitting constraints.
Publication
Authors
M.A. Nichols, A.J.R. Godfrey, G.R. Wood, C.G. Qiao, S. Ganesalingam
Keywords
G X E interaction; incomplete data; cluster analysis; imputation; two-stage method
Online Articles (55)