Resumen

DUARTE VERA, Fernando; SANDOVAL CASTRO, Carlos  y  SARMIENTO FRANCO, Luis. Evaluation of the CNCPS-S Model to Predict the Pelibuey Sheep Growth. Rev. Cient. (Maracaibo) [online]. 2008, vol.18, n.3, pp.296-304. ISSN 0798-2259.

The Cornell Net Carbohydrate and protein System for sheep (CNCPS-S) was evaluated on its prediction capacity for daily weight gain and food intake of hair sheep. A data base containing information from 29 experiments, 1189 Pelibuey sheep and 93 diets was utilized. Data were grouped by sex, model adequacy was analyzed using correlation and lineal regression analysis (r and r²), the mean square error or standard error of the estimate (MSE), the concordance correlation coefficient (CCC, “C_b y P_c “), the mean square error of prediction (MSEP), the model efficiency (MEF), and the determination coefficient of the model (CD). Daily gain prediction was sub estimated by the model. The r, r² and C_b were higher in females than males with values of 0.84, 0.71, 0.87, 0.48, y 0.69, 0.79, respectively, the MEF values indicated a better efficiency with males as compared with females with values of -0.004 y 0.443, respectively; ewes have lower MSE than males, similarly MSEP and CD, the last one closer to the unity, indicated a better prediction by the model. It was concluded that the CNCPS-S model was more accurate for females and more precise for males. The evaluation of the model indicated that the CNCPS-S can be used for predicting daily gain on hair sheep, however some adjustments are needed in order to improve the estimations with hair sheep.

Palabras clave : Hair sheep; daily weight gain; model evaluation of CNCPS-S.