Q$1$:

A$.$ Briefly explain the bias$-$variance trade off concept

B$.$ What is the main purpose of feature selection? Under what scenario do you think performing feature selection would NOT be very helpful to do$?$

C$.$ When using glmnet, what does the penalty term effectively do to your regression model?

D$.$ Suppose that the true relationship between a response variable and one predictor variable is $y$=\backslash$beta$\_0+\backslash$beta$\_1$x$+\backslash$beta$\_2$x$^2+\backslash$varepsilon$

$.$ For a given data set, if the analyst fits a fifth degree polynomial to the data, the validation MSE will be higher or lower than the validation MSE generated from a quadratic model. Answer the question again replacing the word $$validation$$ with $$train$$