CineSLEUTH: Your Ultimate Engine for Movie Recommendations

CineSLEUTH is an intelligent movie recommendation system that combines advanced machine learning techniques such as TF-IDF vectorization, cosine similarity, fuzzy matching, and Apriori-based collaborative filtering. The system provides highly accurate and personalized movie recommendations for users.

---

Features

1. Content-Based Filtering:

   • Uses TF-IDF vectorization to calculate similarity based on genres and tags.
   • Cosine similarity is applied to find the closest matches.
   • Fuzzy matching ensures typos are handled effectively.

2. Collaborative Filtering:

   • Utilizes the Apriori algorithm to process user ratings and generate association rules for better recommendations.
   • Generates frequent itemsets for movies based on user behavior.

3. Interactive User Experience:

   • Auto-complete suggestions for movie titles with clickable options.
   • Streamlit interface for an interactive and responsive application.

4. Real-Time Data Processing:

   • Handles large datasets with chunk-based reading for scalability.

---

Installation

1. Clone the repository:

   git clone https://github.com/AaryanGole26/CineSLEUTH.git
   cd CineSLEUTH

2. Install the required Python libraries:

   pip install -r requirements.txt

3. Add the required CSV files to the repository:

   • movies.csv
   • ratings.csv
   • tags.csv
   • genome_tags.csv

   You can fetch these datasets from

---

Usage

Run the application locally using Streamlit:

streamlit run CineSLEUTH.py

Key Features in Action:

1. Enter a movie title in the input box.
2. Get auto-complete suggestions and click to see recommendations instantly.
3. Analyze the recommendations through a table of similarity scores.
4. View Apriori-based rules generated from collaborative filtering.

---

Output Visualization

Sample Output Video

Here is a sample video showcasing the output of CineSLEUTH:

Example video showing movie recommendation process and visualization.

---

Flowchart of the System

The flowchart below represents the working of CineSLEUTH:

Flowchart depicting the movie recommendation process from user input to final recommendations.

---

File Structure

• CineSLEUTH.py: Main Python script for the Streamlit application.
• movies.csv: Movie metadata.
• ratings.csv: User rating data.
• tags.csv: Tags associated with movies.
• genome_tags.csv: Genome tags dataset.
• requirements.txt: Required Modules.

---

Technologies Used

• Python Libraries:

  • pandas for data manipulation.
  • apyori for association rule mining.
  • sklearn for TF-IDF vectorization and cosine similarity.
  • fuzzywuzzy for fuzzy matching of movie titles.
  • matplotlib and seaborn for data visualization.
  • streamlit for interactive UI.

• Algorithms:

  • Content-based filtering (TF-IDF + Cosine Similarity).
  • Collaborative filtering (Apriori Algorithm).

---

Example Rules Generated (Apriori Algorithm)

• Rule: Movie A → Movie B
  • Support: 0.02
  • Confidence: 0.85
  • Lift: 3.5

---

Contribution

Feel free to fork the repository, raise issues, or contribute to the project. Contributions are always welcome!

---

License

This project is licensed under the MIT License.

---

Contact