ABSTRACT

This study combined exploratory factor analysis (EFA) and Lasso regression to give a deeper insight into production-economic variables affecting short-run dairy cow lifetime profit, DCLP, at the farm level. The study rests on data on heifer growth, feeding, financial, and production variables from 13 farms. We calculated costs and income for an average animal per month and herd. Costs and income were discounted and summarized to a DCLP at time of birth, converted to profit as a monthly annuity equivalent value, MEQ. MEQ was regressed on the 53 original variables (Lasso) or on factor scores (EFA) derived from the original variables. Both EFA and Lasso regressions were used to deal with co-linearity problems. The EFA provides a higher resolution of the underlying quantities than Lasso regression. The factors improving DCLP were reduced roughage costs and high milk yield combined with the lowest possible age at first calving in the data.

摘要

这项研究结合了探索性因素分析（EFA）和Lasso回归，对农场水平上影响短期奶牛终生利润DCLP的生产经济变量提供了更深入的了解。该研究基于来自13个农场的小母牛生长，饲喂，经济和生产变量的数据。我们计算了每个月平均畜群的成本和收入。将成本和收入折现并汇总为出生时的DCLP，然后转换为利润，作为月度年金当量值MEQ。MEQ对53个原始变量（Lasso）或从原始变量得出的因子得分（EFA）进行了回归。EFA和Lasso回归均用于处理共线性问题。与Lasso回归相比，EFA提供了更高的基础数量分辨率。