A try to building the best linear regression model for prediction and controlling the consumption of electric energy In Baghdad

Abstract EN

This paper tries to build the best linear regression model (BLRM) using Least square method (L.S.M) for data taken from random sample of families in Baghdad to predict and control the local consumption of electric energy.

To achieve that aim it has depended on the examination of residuals of linear models.

It's used "SPSS system" for the following: - Detect the outliers and the influential observations of them and also the multicollinearity problem.

- Meet the usual assumptions about the errors (UAE) - Find the mean square errors (MSE) and the mean square predicted errors (MSPE) as criteria to arrive at BLRM.