Estimation of income with using Ridge Regression analysis in layer hen industry

Creative Commons License


ANKARA UNIVERSITESI VETERINER FAKULTESI DERGISI, vol.62, no.1, pp.69-74, 2015 (SCI-Expanded) identifier


The aim of this study was to estimate the sale income using age, survival rate, egg weight and production data in layer hen industry. The egg prices were obtained from Kayseri Tavukculuk Sanayi ye Ticaret A.S. (Kaytas). The performance values of Bovans White hybrid were used in the analysis. The results of the study showed the presence of multi-collinearity problem due to, correlation coefficients between independent variables are close to 1, condition number is over 1000 (1463.5) and variance inflation factor (VIF) values of three independent variables are over 10 (264.7; 259.7 and 10.9). Due to multi-collinearity problem, data were analyzed with ridge regression (RR) method which is alternative to least squares regression (LSR) and compared each other. Income estimation with LSR method before removing the independent variables which are in multi-collinearity with each other the model's coefficient of determination is calculated as R-2 = 0.99; while the R-2 = 0.98 with selected K value with using RR method. In addition; standard errors of the regression coefficients of independent variables were lower in RR method. In conclusion, RR method provided, lower standard errors, more stable, consistent and suitable estimates compared to LSR method.