Question 10 [1 point) You want to predict student performance in an organic chemistry class. For each student, you have data on several potentially relevant predictors such as work ethic, IQ score, ability to visualize in three dimensions, ability to mentally rotate shapes, proactive attitude, the ability to cope with stress as well as the number of labs and lectures they attended. This is a large class, as all of the premeds have to take it, so your sample size is 500. The rows of the dataset [each row represents data from one student) are ordered in the same order as the students sit in the lecture hall, from front to back and left to right. Out of the following options. this one is the best way to model this 0 a} You use all of the predictors, but split the dataset in half, at random. You use one part of the dataset to build a multiple regression model. You then use the coefficients from this model to make the predictions and compute the RMSE in the other part. O b) You use all of the predictors, but split the dataset in half. You use the first half of the dataset to build a multiple regression model. You then use the coefficients from this model to make the predictions and compute the RMSE in the second half of the dataset. This is reasonable because this model effectively incorporates the order information in the model, which is likely to improve accuracy. 0 c] You pick one of the predictors and build a simple linear model, in order to avoid colinearity concerns. 0 d] You use all of the predictors and all of the data to build a multiple regression model. You then use the coefficients from this model to make the predictions and compute the RMSE. This is the most efficient way as this approach uses *all* [not just half!} of the data, so these predictions will be most precise. O e) You pick one of the predictors and one obvious confound and do a partial correlation