Discussion questions#
Other than allowing for classifying cases where Y has more than two groups, what other advantages do LDA & QDA have over standard logistic regression? When would logistic regression be preferred and why?
The lecture went over how standard logistic function defines a state transition in Y. If a logistic regression model fit produces a function with a shallow slope, what does that tell you about the relationship between X & Y? How would you interpret the relationship between specific predictor variables in X (e.g., \(X_1\)) and the state of Y?