- Developed user-based movie recommendation system by implementing user-user collaborative filtering.
- Used Netflix movie dataset containing 100,000 user records for developing recommendation engine.
- Reduced run time and space complexity significantly.
- Implementation in both C++ and Python separately.
- Please open "AnjanaTihaMachineLearningProjectvFinal.ipynb" file in python notebook(Anaconda contains most libraries)
- Provide "movie_titles.txt" and "ratings.txt" file location in each file read function for moie_titles and ratings.
- Run all cells in the file.
- After running, user id and recommendation size K will be asked.
- Upon providing user id and recommendation size, K titles and year of recommended movies will be displayed from most most recommended in descending order.
- Please open "movie_recom.cpp" file in a isual Studio/Eclipse/ other C++ software tools project. Use an ecplipse project already provided under the zip file(Contains two seperate implementation for C++ and Python). Please use 7zip software for unzipping "http://www.7-zip.org/download.html"
- Provide "movie_titles.txt" and "ratings.txt" file location in argument in Visual Studio/Eclipse/ other C++ software tools in source code.
- Build and Run.
- After running, user id and recommendation size K will be asked.
- Upon providing user id and recommendation size, K titles and year of recommended movies will be displayed from most most recommended in descending order. Note: C++ implementation used C++ 11 version.
Current Version : v1.0.0.0
Last Update : 04.30.2017