🎬 Movie Recommendation System (Cosine Similarity Based)

This project is a simple movie recommender system built using item-based collaborative filtering with cosine similarity. It suggests movies similar to a given movie based on user rating patterns—no machine learning models used.

---

🚀 Project Overview

• Approach: Item-based collaborative filtering using cosine similarity
• Dataset: movies.csv (https://www.kaggle.com/datasets/harshshinde8/movies-csv)
• Tech stack: Python, pandas, NumPy, Seaborn, Matplotlib

---

🧬 Data Description

• movies.csv: Contains movieId, title, and genres

---

🧹 Preprocessing Steps

• Read and cleaned the dataset
• Created a pivot table: rows = users, columns = movies, values = ratings
• Replaced missing values with empty string
• Computed pairwise cosine similarity between movie columns

---

🤖 Recommendation Logic

• Input a movie title
• Extract its column vector from the user-movie matrix
• Calculate cosine similarity with all other movies
• Return the top N most similar movies (excluding the input)

---

💻 How to Run

• Clone this repo
• Install dependencies:

• pip install -r requirements.txt
  

Run the notebook:

• jupyter notebook MovieRecommendation.ipynb

---

📽️ Sample Output

• Input Movie: Forrest Gump
  • Top Recommendations:
  • The Shawshank Redemption
  • Apollo 13
  • Piladelphia
  • Rain Man

---

🧠 Limitations

• Only recommends based on rating patterns, not content or genre
• No ML model or deep personalization for individual users