第三章多元线性回归2016.pptVIP

3、the relationship between F and R2 R2 =0，F=0 R2 =1，F=∞ Just as the R2, F shows the goodness of fit of regression model, namely how the multiple explanatory variables infect the dependent variable 3.5 prediction with multiple regression The method of prediction with multiple regression is just the same as with two-variable regression 3.6 the Extension of Testing The purpose: to verify the economic theory, to select the more fitted model. To test the equality of two regression coefficent H0: 2. The general F testing Set null hypothesis: H0: the m restrictions Run the unrestricted regression and restricted regression respectively. compute the F statistics 4. Decision rule: If ，reject null hypothesis, otherwise we can not reject it Simple discussion How to evaluate a model? How to chose a model? Chapter ThreeMULTIPLE LINEAR REGRESSION ANALYSIS Main points 1.Prf and srf 2.Classical assumption 3.OLS estimation method 4. OLS estimator’s statistical property 5.Adusted R square 6. T test and F test 7. Prediction 8. The extension of regular test 3.1 multiple linear regression model and basic assumption3.1.1 multiple regression model Y is dependent variable, X1、X2…XK are independent variable . u is disturbance Β1β2…..Βk are partial regression coefficient k is the number of explanatory variables. a simple example of multiple regression model, three-variable regression model PRF: Yt= ?0+ ?1X1t+?2X2t +ut E(Yt)= ?0+ ?1X1t+?2X2t SRF Attention: meaning of regression coefficient： meaning of partial regression coefficient: Measure the change in the mean value Y,per unit change in its varible, holding other variables constant. 3.1.2 BASIC Assumption ASSUMPTION 1 Model is linear in parameters，X is fixed-value, Or zero covariance between ui and Xi Because X is fixed value and u is random variable, this assumption is satisfied automatically. This assumption