1. Which of the following statements is/are true for LASSO? (There might be more than 1 answer)
   1. Restrictions of LASSO are usually set within the training sample performance
   2. The more restrictive we are, the worse LASSO performs (in the sense of MSE)
   3. LASSO is based on restriction connected to the absolute values of the estimated betas
   4. For classification problems, LASSO is equivalent to logit if the restrictions are not set
   5. LASSO is standardly used for large datasets to avoid heteroskedasticity