Tasks Software is a software catalog firm that sells games and Educational software . It started out as a Bothare manufacturer and then added third - party . title : to its offerings. It recently revised its collection of items in a new catalog, which it mailed out to it's customers . This mailing yielded 1010 purchases. Eased on these data, Izj ko wants to devise a model for predicting the spending amount that a purchasing customer will yield . The file Tajkacz contains information on 1010 purchases . The table below describes the variables to be LEEd in the problem ( the car file contains additional variables that You can Ignore for this problem'! FRED Number of transactions In the preceding year LAST UPDATE Number of days. since last update to customer record* WEB GENDER Whether customer purchased by Web ander at least once Male or female* ADDRESS RES Whether it is a residential address ADDRESS US` Whether it is a U. S. Address SPENDING [response! Amount Spent by customer in test mailing (in dollars ! For the next set of questions , partition the Idad records into (```; training and (4^; validation Eats . a. Fit a multiple linear regression model for GEE _LING using all six predictors . What is the Estimated predictive Equation, and what does it tell you about the type of purchasEr that is most like` to spend a LATE = amount of money !" 6. Compare the model above with the models that result from forward selection , backward elimination . and stepnice variable selection procedure . c. Evaluate the predictive accuracy of the above models by examining their performance on the `Elidation set