A New Suggested Method for the Selection of Best Regression Equation Depending on the Partial Correlation coefficient

Abstract

This research suggested a new heuristic method to select the best regression equation. Assuming that we have a set of predictor variables. The choice of some of those variables is required to obtain the information that we get it if we used all those variables. The research presents the used methods to select the best regression equation also presents a new method that depends on the partial correlation coefficient. This paper compares the new method with stepwise regression procedure by using real data depending on criteria mean square error (mse).The new method showed ease of use and high efficiency.