How samarthpardhi/mixture-models works
Project type and purpose
• This is an open-source, Python-based mixture-model library and command-line toolset for heterogeneous data clustering (Gaussian, categorical, hybrid) with feature-selection, accompanied by Jupyter notebook examples.
File structure
• Top level:
• mixtureModels.py – core EM-based model classes
• learn_*.py & learn_scikit_gmm.py – CLI entry points for each model
• utils.py, plotting.py, cluster_stats_new.py, evaluate.py – helper modules for data I/O, visualization, and evaluation
• exampleOut, outCat, outGauss, outScikitGMM – directories capturing model outputs
• examples/ – Jupyter notebooks demonstrating usage and results
• No external build or deployment scripts—this is a library/tool with CLI and notebook interfaces.
README for additional insights
• Confirms the core models (general-covariance GMM, diagonal GMM, categorical, hybrid)
• Highlights command-line arguments, dependencies (NumPy, SciPy, scikit-learn, Matplotlib, seaborn)