Open AccessPrincipal component regression of the morphostructural traits of West African dwarf sheep
Author(s) -
Abdulraheem Adedeji Bello-Ibiyemi,
Abdulraheem Adedeji Bello-Ibiyemi,
Mathew Wheto,
A. S. Adenaike,
John S. Decampos,
D. O. Ogunlakin,
M. Atunnise,
S. Shola,
C. O. N. Ikeobi
Publication year - 2021
Publication title -
nigerian journal of animal production
Language(s) - English
Resource type - Journals
ISSN - 0331-2062
DOI - 10.51791/njap.v43i2.963
Subject(s) - rump , multicollinearity , variance inflation factor , girth (graph theory) , principal component analysis , flock , varimax rotation , biology , mathematics , statistics , linear regression , best linear unbiased prediction , collinearity , crossbreed , zoology , veterinary medicine , selection (genetic algorithm) , medicine , ecology , cronbach's alpha , combinatorics , artificial intelligence , descriptive statistics , computer science
Evaluation of body weight and seven morphostructural traits (body length, wither's height, rump height, fore cannon bone length, chest depth, hip width and hearth girth) of 200 West African dwarf (WAD) sheep of age range 0-1, 1-2, 2-3 and 3-4 was used in determining age and sex effects using General linear model procedure. The sheep were traditionally managed in flocks of southern and north central Nigeria. Interdependence magnitude (redundancy) was estimated to be positive and significant through pearson's correlation estimates between body weight and linear body dimensions (r = 0.04 – 0.81; P < 0.01). Among the linear type traits, highest correlation was observed between body weight and hearth girth in males (r = 0.85) and females (r = 0.81) respectively. Multicollinearity contrast at variance inflation factor (VIF) threshold 3.0 was found to be most associated with rump height, hearth girth and body length; with low tolerance modules. Collinearity effects were further orchestrated with computations of eigen values, condition indexes and variance proportions. The component solution through VARIMAX orthogonal rotation generated one principal component for WAD sheep of age range 0-2 years, and two principal components for age range 2-4 years. The principal component based on regression models revealed, body weight was best predicted from heart girth, and combination of rump height and wither's height measurements.