09_ml_classification

Machine learning: classification

By Gianluca Campanella (g.campanella@estimand.com)

Objectives

By the end of the session, you should be able to:

• Describe some of the metrics used for (binary) classification
• Apply cost-benefit analysis to classification problems
• Use scikit-learn to fit cross-validated, regularised k-NN and logistic regression models

Plan

The session is designed to be delivered over three hours (including breaks).

Topic	Time
Introduction to classification	30 minutes
k-NN classifier using scikit-learn	30 minutes
Logistic regression using scikit-learn	30 minutes
Exercises	60 minutes

Materials

• Introduction to classification
• k-NN classifier using scikit-learn
• Logistic regression using scikit-learn
• Exercises and solutions