Logistic Regression (page 6 of 6)

A few advantages of this algorithm are: 1) it doesn't require high computation power, 2) is easy to implement, 3) is easily interpretable, and 4) it doesn't require scaling of features. A few disadvantages of this algorithm are: 1) it is not able to handle efficiently a large number of categorical features, 2) is vulnerable to overfitting without specifying the regularization hyper-parameter, and 3) it typically performs poorly with features that are not highly correlated to the target.



Reading Check: Having positive & people around you keeps you happier. InClass Example: Predicting Type 2 Diabetes