Logistic Regression

Questions:

1. How effective is the linear regression approach to the problem?
2. Task 1: Compare and contrast L2 Loss vs LogLoss.
3. Task 2: Explain how effective logistic regression is compared with linear regression.
4. Task 3: Here, just report the best values you can achieve for AUC/accuracy and what hyperparameters you used to get them.

Save your answers in lab09_1.txt.

Sparsity and L1 Regularization

Questions:

1. Why are we regularizing with respect to sparsity?
2. How does L1 regularization increase sparsity?
3. Task 1: Here, just report the best log loss value / model size you can get and what gamma value you used to get them.

Save your answers in lab09_2.txt.

Classifying movie reviews: a binary classification example^* — Get this code to run and then answer the following questions.

1. Try Chollet’s “Further experiments”. Do any of the alternatives do better than the suggested architecture? Why or why not?

Save your answers in lab09_3.txt.