Movie Recommendation System is a simple thus fully functional Data Mining project, that showcases the process from which all of the raw data pass through before they are being finally used and manipluated, in order to provide some handy information considering the suggestion of movies, based on some given preferences. The aforementioned preferences can be either some movie / TV show titles themselves or some descriptions that best describe a viewer's taste.
This project was developed in order to help in a study done on the movies of the well-known platform Netflix.
The project is seperated into four (4) main parts:
- Data Preprocessing (missing data handling, lemmatization).
- Studying the data and extracting statistical data, in order to completely comprehend the provided datasets.
- Implementation of the recommendation system (used Bow and TF-IDF models combined with Jaccard-Tanimoto coefficient and cosine similarity).
- Running the program and displaying the final results.
The project's data consist of three (3) .csv files (netflix_titles.csv
, IMDb movies.csv
, IMDb ratings.csv
) that can be found under Movie-Recommendation-System/data/
directory in the project's repository and contain:
- Movie and TV Show titles from Netflix.
- Movie information from IMDb.
- Movie ratings from IMDb.