MovieRecommendationSystem

This project explores various machine learning techniques to develop a comprehensive movie recommendation system. We integrate content-based, collaborative filtering, and hybrid models to offer personalized movie recommendations tailored to individual user preferences.

Team Members :

• Aditya Sahani(B22CS003)
• Raunak Singh(B22CS085)
• Arjun Bhattad(B22AI051)
• Krishna Chaudhary(B22EE090)

Demo Website :

Project Website :

Report : PRML_PROJECT_REPORT

Dataset : Dataset Link

Abstract

Our recommendation system leverages state-of-the-art machine learning and natural language processing algorithms to analyze movie features such as genre, director, and synopsis. Through extensive experimentation and evaluation, we optimize the system's performance considering metrics such as accuracy, diversity, and scalability. The deployed system offers a user-friendly interface for discovering relevant movies based on user preferences and browsing history.

Table of Contents

• Introduction
• Objectives
• Approaches Tried
• Experiments and Results
• Results

Introduction

In today’s digital age, recommendation systems play a crucial role in helping users navigate the overwhelming abundance of content. In the realm of movies, our project focuses on developing a recommendation system that provides personalized suggestions based on individual preferences.

Approaches Tried

Content-Based Filtering

We analyze movie attributes such as genre, cast, and director to recommend similar movies based on their content characteristics. Techniques like cosine similarity and Jaccard similarity are employed for this purpose.

Collaborative Filtering

We recommend items based on the preferences of other users, leveraging user-item interactions to identify users with similar tastes and preferences. Methods such as matrix factorization and nearest neighbors are utilized in collaborative filtering.

Hybrid Filtering

Our hybrid model combines the strengths of both content-based and collaborative filtering methods to enhance recommendation accuracy and coverage. Techniques like cosine similarity and matrix factorization are integrated to provide comprehensive recommendations.

Experiments and Results

Experimentation

We experiment with various algorithms and techniques, including cosine similarity, matrix factorization, and KNN, evaluating their effectiveness in recommending movies.

Results

Results indicate the performance of different recommendation approaches, highlighting their advantages and limitations.

Summary

Our project contributes to the advancement of recommendation systems by demonstrating the effectiveness of integrating multiple techniques to deliver personalized and engaging movie recommendations in real-world applications.

References

Ref1 : https://www.youtube.com/watch?v=ZspR5PZemcs

Ref2 : https://www.researchgate.net/publication/332579725_Collaborative_Filtering_Recommendation_Model_Based_on_k-means_Clustering

Ref3 : https://medium.com/@jwu2/improving-collaborative-filtering-with-clustering-88c63bdae7cc

Ref4 : https://towardsdatascience.com/how-to-build-a-movie-recommendation-system-67e321339109

Ref5 : https://medium.com/sciforce/deep-learning-based-recommender-systems-b61a5ddd5456

Ref6 : https://arxiv.org/pdf/1707.07435.pdf