Factor analysis and its application to carcas data of sheeps

Abstract

Most of agricultural experiments allow collecting multiples phenotypes from each experimental unit. Univariate analysis method, which evaluate each phenotype separately are limited in such a case. Consequently, multivariate analysis methods that allow analysis and interpretation of results of all phenotypes together are employed.

In this study, factor analysis, which is a multivariate technique, is described and its application possibilities in agriculture is evaluated. Interpretation of results were addition its application and shown on a data set. For this purpose; 46 lamb breeds from different were used, and carcas measurements of head weight, omental and mesenteric fat, testis weight, kidney and pelvic weight, neck weight, tail weight, heart-lungs-liver weight, spleen weight, leg weight, shoulder weight, flank weight, back-loin weight, cutlet sample, lean weight and bone weight were phenotypes which were used in this study. Factor analysis were performed by using Minitab’s factor analysis menu and algebraic calculation via matrix notation was performed by Minitab package program. Result of factor analysis application on to the data set shows that 3 factor can explain 94 % of the total variation of the original phenotypes. First factor was a combination of  head weight, omental and mesenteric fat, testis weight, kidney and pelvic weight, neck weight and tail weight, second factor heart+lungs+liver weight, spleen weight, leg weight, shoulder weight and flank weight and third factor back-loin weight, cutlet sample, lean weight and bone weight