How khanhnamle1994/movielens works
Project type and purpose
• This is a data-science demonstration project showcasing four movie recommendation engines (Content-Based & Collaborative Filtering, SVD, Deep Learning) built against the MovieLens 1M dataset.
• It is organized as Jupyter Notebooks plus a utility script (CFModel.py), data files, and saved model weights.
File structure
• Top level:
• Notebooks: Data_Processing.ipynb, Content_Based_and_Collaborative_Filtering_Models.ipynb, SVD_Model.ipynb, Deep_Learning_Model.ipynb
• Utility script: CFModel.py (likely encapsulates shared functions or runs experiments in script form)
• Raw data folder “dat/”: users.dat, movies.dat, ratings.dat
• Derived CSVs: users.csv, movies.csv, ratings.csv
• Model artifact: weights.h5 (deep model weights)
• No separate frontend/backend—this is a batch‐style ML pipeline.
Key insights from README
• Four distinct modeling approaches, each in its own notebook, all operating on the same preprocessed data.
• Dependencies reflect data handling (pandas, numpy), classical ML (scikit-learn, surprise), deep learning (keras, h5py), and visualization (matplotlib, seaborn, wordcloud).