Expectation-Maximization Algorithm

Algorithm Control

E-Step (Expectations) M-Step (Maximization) Full EM Iteration Run Until Convergence

Data Generation

Preset Dataset Three Clusters Two Overlapping Elongated Clusters Unbalanced Sizes Generate Data

Parameters

Number of Clusters (K): 3 Points per Cluster: 50 Convergence Threshold: 0.001

Randomize Initial Gaussians Clear All