Question 1 1 pts A digital marketing rm is building a regression model to predict whether a given tweet has a political slant (y=1) or not is fundamentally not political in nature (y=0). Their data scientists assemble a large data set of tweets that have been tagged as having a slant or not. The team has engineered many useful features from the text of each tweet (e.g. whether the name of a political party is mentioned, whether a political candidate is mentioned, time of day, et al). They have thousands of possible features that could be included in a model. They propose to use stepwise selection to nd an appropriate set of features to include in the model. Which of the following would be appropriate advice for this data science team? Choose the best answer. C] Simply use exhaustive enumeration, tting all possible models under consideration until all combinations of features and their interactions have been considered in turn. C] Build a model with as many features as are available, along with their pairwise interactions. [:] Stepwise selection is an appropriate tool here, because it is not known which features should be included in the regression model. C] Stepwise selection is an appropriate tool here, because interpreting the model's coefcients and condence intervals is more important than model performance. [:] Stepwise selection is appropriate for isolating partial relationships, but it is not appropriate for building predictive models