Glossary term
Glossary term
Architecture
A model that predicts labels from a set of one or more features. More formally, discriminative models define the conditional probability of an output given the features and weights; that is:
p(output | features, weights)
For example, a model that predicts whether an email is spam from features and weights is a discriminative model.
The vast majority of supervised learning models, including classification and regression models, are discriminative models.
Contrast with generative model.
For example, a model that predicts whether an email is spam from features and weights is a discriminative model.
The vast majority of supervised learning models, including classification and regression models, are discriminative models.
Contrast with generative model.
Created for this library
A bank trains a discriminative model to estimate the probability of default directly from applicant features, instead of modeling the joint distribution.
A search-quality team uses a discriminative model that scores query-document pairs as relevant or not relevant.
A churn team uses a discriminative model to estimate the probability of cancellation conditional on customer features in the next 30 days.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License