Yelp Restaurants Recommendation System

Welcome to the Yelp Restaurants Recommendation System project! This system aims to provide personalized recommendations to users and predict user-business pairs based on their preferences and behavior using data mining techniques, Spark RDD, and collaborative filtering algorithms.

Website

https://yelprecommendationsystem.wl.r.appspot.com/

Note

The files app.py and RecommendationSystem.py, which house the primary Spark code and application logic, are securely stored within a private repository in accordance with the project requirements. This project serves as an extension of a previous course project, facilitating user interaction through a user-friendly interface. Users can conveniently select users and businesses, enabling them to view recommendation scores with ease

Demo Video

Recommendation_System.mp4

Features

• XGBoost-Regressor based Recommendation: Utilizes XGBoost (XGB) as the primary recommendation engine.
• Collaborative Filtering: Utilizes item-based collaborative filtering with Pearson Similarity to provide accurate recommendations by analyzing item similarities.
• Hybrid Recommendation System: Enhances recommendations through feature mining, achieving significant RMSE reduction from 1.09 to 0.9798.

Techniques Used

• Data Mining: Extracts insights and patterns from very large Yelp dataset.
• Spark RDD: Utilizes Spark Resilient Distributed Datasets for parallel processing and efficient data manipulation.
• Collaborative Filtering: recommendations are made by identifying businesses that are similar to the businesses the user has interacted with or shown interest in.

Results

The recommendation system demonstrates promising performance in predicting user-business ratings, with an RMSE reduction achieved through feature mining. The hybrid approach combining collaborative filtering and feature engineering enhances the accuracy of recommendations.

Usage

1. Setup: Ensure you have Spark installed and configured in your environment.
2. Data Preparation: Load the Yelp restaurant dataset into your environment.
3. Execution: Integrate and execute the provided Spark code to generate recommendations.
4. Web Interface: Utilize app.py (stored in the private repository) to integrate the system into a web interface for user interaction.

Contributors

• Sharvari Kalgutkar