Welcome to the comprehensive repository packed with meticulously designed notebooks to showcase a variety of advanced methodologies and techniques for building high-performing recommendation systems.
Structure and Features The repository houses two primary folders:
-
recommendersys/filtering-correlation-popularity/src/: This folder contains a series of notebooks that demonstrate different recommendation system techniques such as content-based filtering, K-Nearest Neighbors (KNN)-based collaborative filtering, correlation-based systems, and popularity-based techniques. A unique data collection notebook from TMDB is also part of this directory, providing firsthand experience in acquiring datasets for such systems.
Files included:
- Content_based_recc_system.ipynb
- KNN_based_Coll_filtering.ipynb
- Recc_system_co_relation.ipynb
- avg_weighted_popularity_based_technique.ipynb
- data_collection_tmdb.ipynb
-
recommendersys/content-based/: This folder is dedicated to a content-based recommendation system with a specific focus on anime recommendations.
Files included:
- 00_anime-content-based-recommendation-system.ipynb
Additional directories for ipynb_checkpoints and data, which include folder structure, complex notebooks, and extensive datasets, have been added respectively within these main folders.
Future Expansion The work does not stop here. I plan to expand the repository by integrating sophisticated implementations such as Matrix Factorization and ObjectToVec, enhancing the recommendation system's quality and performance.
Requirements To optimally benefit from these elaborate notebooks, understanding and experience in Python, machine learning techniques, popularity-based filtering, content-based filtering, and collaborative filtering are recommended.
Stay tuned.