COMPARISON OF LINEAR AND POLYNOMIAL GROWTH MODELS FOR PREDICTING LIVE WEIGHTS OF TWO RABBIT BREEDS
DOI:
https://doi.org/10.51791/njap.vi.4265Keywords:
Rabbits, Growth models, Model parameters, goodness of fit, weight predictionAbstract
This study investigated the goodness of fit of linear, quadratic and polynomial mathematical functions on weight-age data of Chinchilla and New Zealand white rabbits with a view to predict their weights at various ages. A total of 20 weaner rabbits comprising 10 per breed of New Zealand White (NZW) and Chinchilla (CHA) in a sex ratio of 1:1 were used for the study. Weekly body weight of rabbit was recorded from week 7 to 26The weight-age data were fitted to the growth functions using Levenberg-Marquardt method of Statistical Package for Social Sciences. Goodness of fit of the models was assessed using coefficient of determination (R2 ) and mean square error (MSE). Growth model parameters A, B and C were found to range from 920.4 (Linear model in CHA) - 1006.4(Linear model in NZW), 6.7 (inverse polynomial model in CHA) - 24.2 (Quadratic model in CHA) and-0.097 (Quadratic model in CHA) - 0.194 (inverse polynomial model in CHA) respectively. Based on R2 and MSE, the Quadratic model (R2=0.97, MSE=4082.79) and Inverse polynomial model (R2=0.97, MSE=3051.27) were better fit for prediction of body weight than the Linear model (R2=0.85, MSE=24606.60). In conclusion, Inverse polynomial or Quadratic model fit description of weight-age relationship in rabbits
