Gaussian Mixture Model

A probabilistic model that represents data as a weighted sum of multiple Gaussian distributions to enable soft clustering.

GMMs provide a probabilistic framework for clustering by assuming data points originate from a finite mixture of Gaussian distributions. Unlike K-Means (which uses hard assignments), GMMs utilize soft clustering: each point receives a probability score for every cluster. The model optimizes parameters via the Expectation-Maximization (EM) algorithm to define cluster shapes (spherical, diagonal, or full covariance). This flexibility makes GMMs ideal for complex tasks like biometric voice identification or detecting anomalies in financial datasets where data density varies significantly.

https://scikit-learn.org/stable/modules/mixture.html