Exercise

Imbalanced class metrics

Class imbalance is something that can hamper your model's performance in any machine learning context. This is especially relevant in a machine learning interview if you are asked what to do if you are given a dataset with an imbalanced class, as some data is imbalanced by design such as insurance fraud data.

In this exercise you'll use sklearn to create a logistic regression model and print the confusion matrix along with several evaluation metrics to get a better understanding of how to interpret Machine Learning models from datasets that have a class imbalance.

Recall the class imbalance you saw previously in loan_data. The number of observations with Loan Status of Fully Paid far outweighs those that are Charged Off:

Class imbalance

Instructions 1/4

undefined XP
    1
    2
    3
    4
  • Import the necessary modules to create a logistic regression model as well as confusion matrix, accuracy, precision, recall, and F1-scores.