EM Algorithm for Finite Gaussian Mixture Models

This repository provides the complete source code implementation of the Expectation-Maximization (EM) algorithm for finite Gaussian mixture models, along with a detailed experimental report and runtime results. The experimental report discusses key implementation aspects such as parameter initialization strategies, convergence criteria, and the iterative E-step (expectation) and M-step (maximization) procedures. We include comparative analysis with other clustering algorithms like K-means, highlighting differences in probabilistic modeling and convergence behavior. The implementation features visualization of runtime results through Gaussian component plotting and probability density function displays, enabling better understanding of algorithm performance and cluster separation quality. Additionally, we discuss practical applications of the algorithm, including parameter tuning techniques for optimal results in real-world scenarios, such as selecting the appropriate number of mixture components using information criteria (AIC/BIC). This comprehensive resource aims to facilitate deeper understanding and effective implementation of Gaussian mixture models for various machine learning applications.