- In this section you learned about a different supervised learning technique: classification! Specifically, you practiced building a very basic classification model from scratch - a logistic regression model
- Logistic regression uses a sigmoid function which helps to plot an "s"-like curve that enables a linear function to act as a binary classifier
- You can evaluate logistic regression models using some combination of precision, recall, and accuracy
- A confusion matrix is another common way to visualize the performance of a classification model
- Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC) can be used to help determine the best precision-recall tradeoff for a given classifier
- Class weights, under/oversampling, and SMOTE can be used to deal with class imbalance problems
learn-co-students / dsc-logistic-regression-section-recap-online-ds-grad Goto Github PK
View Code? Open in Web Editor NEWLicense: Other