期刊名称：International Journal of Statistics and Applications
印刷版ISSN：2168-5193
电子版ISSN：2168-5215
出版年度：2018
卷号：8
期号：3
页码：129-132
DOI：10.5923/j.statistics.20180803.03
语种：English
出版社：Scientific & Academic Publishing Co.
摘要：Background & objectives: This work was performed to apply different parametric survival models instead of nonparametric ones for predicting pregnancy in Friesian dairy cattle by using days open as a time variable. The models used were exponential, normal, log-normal, Weibull, logistic, log logistic and smallest extreme value. Data for present study were obtained from animal records belongs to different Dakhalia farms (n = 1842) Covering the period between 2009 and 2011. Variables included in this study were days open as a dependent variable and the independent variables were: age at calving, dry period, calving interval, season, and lactation order. The survival time data were modeled by using life data regression procedure of Statgraphics statistical package program. Model parameters were estimated and comparisons among models were done based on Akaike Information Criterion (AIC). The Weibull model was the best option for evaluating time till occurrence of pregnancy where it had the smallest (AIC) value. It also showed a good fit with the studied data. This Weibull survival model can be used to predict the length of time from calving to conception in Friesian. But the Logistic was not appropriate to describe the dataset.
关键词：Days open; Survival analysis; Parametric models; Censored data and Akaike Information Criterion (AIC)
