STATISTIC ANALYSIS OF MULTIPLE CURVILINEAR REGRESSIONS OF FOUR VARIABLES REGARDING THE EVOLUTION OF THE AVERAGE WEEKLY GAIN AT BROILERS

Abstract

Depending on four controlable variables used in broilers nutrition: E (energy), P (protein), L(lysine), M (metyonine+ cystine) have been deduced mathematically multiple curvilinear regressions showing the evolution of corporal mass during entire growth period. In this paper, using these regressions, we determine the average weekly gain of corporal mass. We test using dispersional analysis if there are significant differences between N.R.C. 1994 and the values given by regressions. Using correlation report we decide which of these regressions is optimum

Keywords

• multiple curvilinear regressions
• dispersional analysis
• correlation report