TatevKaren/data-science-popular-algorithms

Implements collaborative filtering for movie recommendations using item-based nearest-neighbor matching on the MovieLens 20M dataset, alongside foundational algorithms like LDA for classification, K-means for unsupervised clustering, and decision trees for interpretable predictions. The repository pairs theoretical papers with multi-language implementations (Python, R, Scala) and includes a novel Cluster Dynamics algorithm that predicts customer migration between segments based on probabilistic class distributions. Each module combines mathematical foundations with practical case studies and step-by-step evaluation methodologies across recommendation systems, dimensionality reduction, and segmentation tasks.

134 stars. No commits in the last 6 months.