Hybrid Recommendation System Project Overview:
-
User-Based Recommendation:
- Analyzes preferences of similar users based on movies watched by a specific user.
- Recommends movies liked by similar users.
- Steps involved:
- Preparation of data by creating a user-movie dataframe.
- Determining movies watched by the target user.
- Accessing data and IDs of other users who watched the same movies.
- Identifying the most similar users to the target user.
- Calculating weighted ratings and recommendation scores.
- Providing top movie recommendations based on scores.
-
Item-Based Recommendation:
- Suggests similar movies based on the characteristics of a particular film.
- Steps involved:
- Extracting the last highly rated movie by the user.
- Preprocessing data including titles, genres, and timestamps.
- Creating a user-movie dataframe.
- Calculating correlation between movies.
- Sorting and selecting top similar movies.
- Providing top movie recommendations.
-
Hybrid Recommendation:
- The hybrid system combines user-based and item-based recommendation methods.
- User-based method relies on user similarity to recommend movies.
- Item-based method suggests similar movies based on characteristics of a highly rated movie.
- The system aims to provide a diverse and personalized recommendation experience.