Regression analysis is the analysis of learning how to build a functional model of the data to be able to explain or predict a natural phenomenon the basis of other phenomena. Analysis is a relationship between two or more variables that are generally expressed in mathematical equations. In a statistics regression model can be said to be good or suitable, if meet with ideal assumptions (classics), namely the absence of autocorrelation, heteroscedasticity and multicollinearity. So that the control of the model needs to be done to examine whether these assumptions be met.