5. The following analyses are based on a dataset with 4 variables, Y, X1, X2 and...

Transcribed Image Text:

5. The following analyses are based on a dataset with 4 variables, Y, X1, X2 and X3. Note C(P) is a measure of bias in a model (and bias is bad...) The expected value of C(P) is the value of p for a given model. The more C(P) differs from p the more biased the model is. Model R2 22 Y=X1 .39 Y=X2 .10 Y=X3 .29 Y=X1 X2 .49 Y=X2 X3 .39 Y=X1 X3 .58 Y=X1 X2 X3 .68 a) Which model would you pick as the best model? Why? (10) b) Is there any correlation between the x variables? (5) 6. Are there any problems with the following residual plots? 5. The following analyses are based on a dataset with 4 variables, Y, X1, X2 and X3. Note C(P) is a measure of bias in a model (and bias is bad...) The expected value of C(P) is the value of p for a given model. The more C(P) differs from p the more biased the model is. Model R2 22 Y=X1 .39 Y=X2 .10 Y=X3 .29 Y=X1 X2 .49 Y=X2 X3 .39 Y=X1 X3 .58 Y=X1 X2 X3 .68 a) Which model would you pick as the best model? Why? (10) b) Is there any correlation between the x variables? (5) 6. Are there any problems with the following residual plots?