Glossary term
Glossary term
Foundations
In supervised learning, a classification problem in which the dataset contains more than two classes of labels. For example, the labels in the Iris dataset must be one of the following three classes:
Iris setosa
Iris virginica
Iris versicolor
A model trained on the Iris dataset that predicts Iris type on new examples is performing multi-class classification.
In contrast, classification problems that distinguish between exactly two classes are binary classification models. For example, an email model that predicts either spam or not spam is a binary classification model.
In clustering problems, multi-class classification refers to more than two clusters.
See Neural networks: Multi-class classification in Machine Learning Crash Course for more information.
Created for this library
A retail support team uses multi-class classification to route tickets into one of 20 product-area queues.
A bank's marketing team uses multi-class classification to assign each customer to one of several lifecycle stages for outreach.
An NLP team uses multi-class classification on news articles to assign each article to one of several editorial desks.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License