Premium
A Shifted Multiplicative Model Fusion Method for Grouping Environments without Cultivar Rank Change
Author(s) -
Crossa José,
Cornelius Paul L.,
Sayre Ken,
OrtizMonasterio R.J. Iván
Publication year - 1995
Publication title -
crop science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.76
H-Index - 147
eISSN - 1435-0653
pISSN - 0011-183X
DOI - 10.2135/cropsci1995.0011183x003500010010x
Subject(s) - cultivar , rank (graph theory) , multiplicative function , fusion , biology , cluster (spacecraft) , gene–environment interaction , function (biology) , mathematics , biological system , statistics , computer science , combinatorics , botany , genotype , genetics , mathematical analysis , linguistics , philosophy , gene , programming language
Genotype × environment interactions of most concern to plant breeders involve cultivar rank change across environments, i.e., crossover interaction (COl). When COls are present, cluster strategies can be used to group environments without significant changes in cultivar ranks. Several methods for classifying environments without COI have recently been proposed. In this study, the shifted multiplicative model (SHMM) fusion method based on a building block principle is used for grouping environments which rank cultivars similarly. The SHMM fusion selects the best action by computing for each new cluster formed, a new set of distance measures with the other clusters and with the unclustered individuals. Experimental data were collected on 55 cultivars evaluated under five irrigation levels in each of two trials (I0 environments). Groups of environments obtained with SHMM fusion were compared with those obtained using cluster analysis based on Euclidean distances computed from unstandardized and standardized data. The SHMM fusion formed four final clusters of environments. The final clusters grouped environments with similar irrigation levels. Conventional cluster analysis based on Euclidean distances produced final clusters with much larger percentage of COls than SHMM fusion. Results show that SHMM fusion is a useful strategy for classifying environments without cultivar rank change.