TY - JOUR AU - Andrew I., Iheagwara AU - Idochi, Okenwe AU - Stephen, O., Ihekuna PY - 2018/08/31 Y2 - 2024/03/28 TI - MODEL SELECTION IN MULTIPLE REGRESSION MODELS USING BUDGETED PROFIT, PRODUCTION AND SALES VARIABLES JF - IJRDO-Journal of Applied Science JA - as VL - 4 IS - 8 SE - Articles DO - 10.53555/as.v4i8.2263 UR - https://ijrdo.org/index.php/as/article/view/2263 SP - 09-20 AB - This study is on model selection in multiple regression models. Data for this study were collected in Nigerian bottling company plc, Owerri plant from 1999 to 2013. The response variable is budgeted profit, while the explanatory variables are budgeted production and budgeted sales. Four regression models; Linear, Lin-Log, Polynomial, and Inverse were examined in this study. The E-views software was used in this study. Four model selection techniques known as; coefficient of determination, Akaike Information Criterion, Schwarz Information Criterion, and Hannan-Quinn Information Criterion were used to select the best model. From the analysis, it can be concluded that the nonlinear models perform better than the linear model. However, in the overall goodness of fit assessment, the study concluded that the polynomial regression model performs far better than the other three regression models used in this study. Therefore, future researchers should look at a similar work by incorporating other nonlinear regression models like Double-Log and Log-Lin Regression models to compare results. It should be noted by future researchers that if Double-Log and Log-Lin Regression models are employed, then Quasi - R2 is needed instead of R2 as employed in this study. ER -