Neocortex 🧠

Search

Search IconIcon to open search

Discriminative vs Generative Models

Last updated Sep 25, 2022 Edit Source

There are two approaches to classification, discriminative and generative models.

# Discriminative Models

Discrimintive models aim to classify objects by just knowing the notes/Posterior Probability. The hard problem that needs to be solved is how we can calculate those probabilities, and to solve it, we need to make strong assumptions.

# Generative Models

Generative models aim to calculate the posterior probability by knowing the Class Prior and the conditional density($p(x|\omega)$). Using the principle: $$ p(\omega|x) \propto p(\omega)p(x|\omega) $$ It requires the class prior and the conditional density to be estimated.


Interactive Graph