Lecture 10: Classification - KNN and Naive Bayes

Content locked

The catalog is public. Reading this content requires approved access.

Introduction

Supervised Machine Learning Approaches

Non-parametric vs. Parametric

Statistical Modeling Approach

K-Nearest Neighbors (KNN)

Algorithm for Classification

Distance Functions

Distance Function Considerations

KNN for Regression

Decision Regions (1-NN)

Effect of k

Maximum Likelihood Estimation (MLE)

Selecting Models

Likelihood Function

Log-Likelihood

Example 1: Coin Flips (Bernoulli Model)

Example 2: Gaussian Model

Classification via MLE

Naive Bayes

Bayesian Classification

Naive Bayes Assumption

Example: Play Outside or Not?

Practice Problems

Key Takeaways