Multivariate binary logistic regression spline analysis on the influence of grass and concentrate composition on cattle pregnancy

Introduction: Estimating pregnancy in cattle is a critical aspect of reproductive management in the livestock industry. This study aims to estimate the likelihood of cattle pregnancy by considering variables such as the amount of grass and concentrate consumed, Body Condition Scoring (BCS), and cattle variety.

Objective: In this study, data from one hundred cattle were analyzed, where each animal was recorded for pregnancy status, grass and concentrate consumption, BCS condition, and variety. The data was then analyzed using the Multivariate Adaptive Regression Splines (MARS) method, which allows for the identification of nonlinear relationships and complex interactions between predictor variables and response. From this analysis, four main estimators were identified, with two showing statistical significance as primary predictors of pregnancy, namely grass and concentrate consumption. 

Results: It was found that the amount of grass and concentrate consumption has an inverse linear relationship with the likelihood of pregnancy, particularly at consumption points of thirty and fifteen kilograms. MARS analysis also showed that BCS and variety play a role in influencing pregnancy, although in this study they were not as influential as feed consumption. Therefore, body condition scoring and the correct selection of cattle variety should also be considered in cattle reproductive management.

Conclusion: This study highlights the importance of proper selection and management of nutrition as a key factor in increasing the likelihood of cattle pregnancy. Through sophisticated statistical analysis, this research provides important insights into effective ways to increase reproductive efficiency in the livestock industry, thereby helping farmers make data-based decisions for animal nutrition management.