Use of non-parametric regression method in livestock

In the study, the non-parametric regression methods of Hodges-Lehmann, Theil-Sen, Mood Brown, and OLS were used on data from 58 Japanese quails to establish the error term, which method to apply, and how to make parameter estimations. When the coefficients in the regression equation were examined, it was found that there were similar coefficients in different methods. When the error term was examined, it can be said that the Hodges-Lehman method had a higher error term and the accuracy in the estimations was relatively higher. Considering the small error term, the low values of the Theil-Sen regression method in most parameters, support the usability of the method. It was found that the Theil-Sen regression method had smaller error terms, especially in the parameters where the independent variable did not show a normal distribution and there were outlier observations. While the OLS estimation method generally had small error term, it was found that especially, when the independent variable was normally distributed it had smaller error terms. The Hodges-Lehmann estimation method gave different, inconsistent results in some parameters with different coefficients and error terms. As a result, in cases when the prerequisites of parametric tests are not met, there are outlier values, and undesirable situations such as multicollinearity, non-parametric regression analysis methods can be considered as an alternative by researchers due to their easy calculation and use.