Cross-Domain Recommendation System

Overview

This project implements a cross-domain recommendation system using datasets from multiple domains (movies, music, and books). The goal is to leverage user interactions across these domains to improve recommendation accuracy.

Project Structure

This project is organized into several key components:

• Data: User-item interaction datasets for movies, music, and books.
• Preprocessing: Cleaning and structuring data for recommendation modeling.
• Modeling: Applying collaborative filtering techniques.
• Evaluation: Assessing the performance of the recommendation system.
• Pipeline: The complete workflow from data preparation to recommendation generation.

Data

The dataset consists of user interaction data across three domains:

• Movies: User ratings for different movies.
• Music: User ratings for music items.
• Books: User ratings for books.

Data Loading

Data is loaded from text files with tab-separated values. The key attributes include:

• user_id: Unique identifier for each user.
• item_id: Unique identifier for each movie, music, or book.
• rating: User-provided rating for the item.

Data Preprocessing

Preprocessing steps include:

• Filtering relevant columns (user_id, item_id, rating).
• Handling missing values to ensure data consistency.
• Merging datasets to prepare cross-domain interactions.

Modeling

A collaborative filtering approach is used for recommendations. The key steps include:

• Building interaction matrices to represent user-item relationships.
• Applying matrix factorization techniques such as Singular Value Decomposition (SVD).
• Generating recommendations based on learned user and item embeddings.

Evaluation

The performance of the recommendation system is assessed using:

• User interaction validation to check recommendation relevance.
• Comparing rating predictions with actual user ratings.

Pipeline

The end-to-end pipeline follows these steps:

1. Load and preprocess data from multiple domains.
2. Train a collaborative filtering model.
3. Generate and evaluate recommendations.

Testing the Model

The trained model is tested with various user interactions to demonstrate its effectiveness.

Example Recommendations:

• User 1: Recommended Movies - [Movie A, Movie B, Movie C]
• User 2: Recommended Music - [Song X, Song Y, Song Z]
• User 3: Recommended Books - [Book P, Book Q, Book R]

Requirements

To run this project, install the following Python libraries:

pip install pandas numpy scikit-learn