Movie Recommendation Systems

This repository contains two different movie recommendation systems:

1. Mood-Based Movie Recommendation System: Recommends movies based on user-selected mood without requiring a specific title input.
2. Title-Based Movie Recommendation System: Recommends movies based on user-input movie title and personalized mood selection, leveraging content-based and collaborative filtering methods.

Each recommendation system uses the MovieLens dataset to analyze movie genres, ratings, and user mood, generating movie suggestions tailored to user preferences.

Table of Contents

• General Information
• Requirements
• Dataset
• 1. Mood-Based Movie Recommendation System
  • Overview
  • Usage
• 2. Title-Based Movie Recommendation System
  • Overview
  • Usage
• Project Structure
• Acknowledgments

---

General Information

This project demonstrates two distinct approaches to movie recommendations:

• Mood-based approach: Generates recommendations based on predefined genre mappings for moods like "happy" or "nostalgic."
• Title-based approach: Combines collaborative and content-based filtering to suggest similar movies based on an input movie title and user mood.

Both systems are implemented in Python and feature a simple user interface using Streamlit.

Requirements

The project requires the following libraries:

• Python 3.7+
• pandas
• numpy
• scikit-learn
• nltk
• streamlit

Install the required packages:

pip install pandas numpy scikit-learn nltk streamlit

Dataset

The MovieLens dataset is used, specifically ml-1m (MovieLens 1M) in .dat format, which includes movie titles, genres, and user ratings.

Loading the Dataset

Due to encoding issues, the dataset is loaded using ISO-8859-1 encoding, with the following files:

• movies.dat: Movie IDs, titles, and genres.
• ratings.dat: User IDs, movie IDs, ratings, and timestamps.

---

1. Mood-Based Movie Recommendation System

Overview

The Mood-Based Movie Recommendation System recommends movies based on a selected mood. Moods such as "happy," "sad," "angry," and "nostalgic" are mapped to genres like "Comedy," "Action," and "Drama." Users select their mood, and the system generates a list of popular movies fitting that mood.

Mood Map

Mood	Genres
Happy	Comedy, Romance
Sad	Drama, Thriller
Angry	Action, Thriller
Fearful	Horror, Thriller
Surprised	Thriller, Mystery
Adventurous	Action, Adventure
Nostalgic	Animation, Family

Usage

1. Run the code in a terminal or editor (e.g., Visual Studio Code):

   streamlit run mood_based_recommendation.py

2. Select a mood from the dropdown and click "Get Recommendations" to see a list of recommended movies.

---

2. Title-Based Movie Recommendation System

Overview

The Title-Based Movie Recommendation System combines content-based filtering (by movie genres) with collaborative filtering (user ratings) to recommend movies based on a selected movie title and mood. The system suggests movies similar to the input title and fitting the selected mood.

Features

1. Content-Based Filtering: Uses movie genres to find movies similar to the input title.
2. Collaborative Filtering: Leverages user ratings with NearestNeighbors to find similar movies based on ratings.
3. Mood-Based Genre Filtering: Filters recommendations by user-selected mood.

Usage

1. Run the code in a terminal or editor (e.g., Visual Studio Code):

   streamlit run title_based_recommendation.py

2. Enter a movie title and select a mood, then click "Get Recommendations" to generate personalized suggestions.

---

Project Structure

├── ml-1m/                          # Dataset folder
│   ├── movies.dat                  # Movies file with movie IDs, titles, genres
│   └── ratings.dat                 # Ratings file with user IDs, movie IDs, ratings, timestamps
├── mood_based_recommendation.py    # Code for mood-based recommendation system
├── title_based_recommendation.py   # Code for title-based recommendation system
└── README.md                       # Project README file

Acknowledgments

• MovieLens Dataset
• Streamlit Documentation