14. Classification

This chapter introduces classification as supervised machine learning for category prediction. You will study nearest neighbors, training and testing, feature tables, implementation, and accuracy.

Chapter Video

Video credit: CrashCourse. (2018, October 31). Supervised Machine Learning: Crash Course Statistics #36. YouTube. Watch on YouTube.

This embedded video is provided for educational reference only. All rights are retained by the respective copyright holder.

Learning Objectives

By the end of this chapter, you will be able to:

1. Distinguish classification from regression and clustering.

2. Use labeled examples to predict categorical outcomes.

3. Explain nearest-neighbor classification with distance and similarity.

4. Separate training data from testing data when evaluating a classifier.

5. Measure classifier accuracy and discuss limits of prediction.

Chapter Flow

Classification

Nearest Neighbors

Training, Testing, and Accuracy

Rows of Tables

Implementing the Classifier

Glossary

#	Term	Explanation
1	Classification	Predicting a categorical label for an observation.
2	Classifier	A model or rule that assigns labels.
3	Label	The category being predicted.
4	Feature	An input variable used for prediction.
5	Training data	Labeled data used to build a classifier.
6	Test data	Labeled data reserved for evaluation.
7	Nearest neighbor	A classifier that predicts from the most similar example.
8	Distance	A numeric measure of dissimilarity between observations.
9	Similarity	How alike two observations are according to selected features.
10	Feature space	The coordinate system formed by model features.
11	Decision boundary	The region where predicted labels change.
12	Accuracy	The fraction of test cases classified correctly.
13	Error rate	The fraction of test cases classified incorrectly.
14	Confusion matrix	A table comparing predicted labels with true labels.
15	False positive	A case incorrectly predicted as belonging to a class.
16	False negative	A case incorrectly predicted as not belonging to a class.
17	Overfitting	When a classifier matches training data too closely.
18	Generalization	Performance on new data beyond the training sample.
19	Supervised learning	Learning from examples with known labels.
20	Validation	Checking model choices using held-out data.
21	Row of a table	One observation or case in a dataset.

Chapter Overview Slides