Listy Biometryczne - Biometrical Letters Vol. 38(2001), No. 1, 33-40
|
WEIGHTED LINEAR JOINT REGRESSION ANALYSIS
Joăo Tiago Mexia1, Dulce Gamito Pereira2, José Baeta3 1Departamento de Matematica, Faculdade de Ci?ncias e Tecnologia, Uiiiversidade Nova de Lisboa, Quinta da Torre, 2825 Monte da Caparica, Portugal 2Departamento de Matematica, Universidade de Évora Colegio Luis Antdnio Verney, Rua Romao Ramăo 59, 7000 Évora, Portugal 3Estagao Agronómica Nacional, 2780 Oeiras, Portugal |
The introduction of weights for the yields of the different cultivars in the different blocks enables us, while carrying out joint regression analysis (JRA), to consider the agricultural significance of the conditions pertaining to the blocks in a network. Moreover it is now possible to use incomplete blocks, thus lighting the local designs and enabling a better coverage of the region under study. Namely, this happens when a designs are used. To carry out the regressions adjustment for weighted linear JRA an algorithm is derived from the sum of sums of weighted residues. Only linear regressions are considered since the use of higher degree regressions does not significantly improve the adjustment.
joint regression analysis, incomplete blocks, randomized blocks, environmental indexes.
