A company C has a VERY complex dataset that is somehow difficult to be classified into two classes. Two data scientists were unhappy that they were told by their manager that they can use ONLY the linear SVM on this complex dataset for their classification task. In an elevator, you overheard the two data scientists were debating the following two questions.

1. What is/are potential reason(s) they must use the linear SVM, instead of the more powerful SVM RBF, for this classification task?

1. If they must use the linear SVM, why not just use the logistic regression that can also produce a linear model?

Before the elevator was closed, can you quickly tell the data scientists your comments for each question in three sentences (i.e. total 6 sentences)? Anything beyond three sentences for each question cannot be heard by the two data scientists, and hence, will **NOT** be graded.