🎥🎬ZEE5 - Recommendation System (Unsupervised Learning)🎬🎥

🎬 Personalized Movie Recommender System

📌 Project Overview

This project focuses on building a personalized movie recommendation system using user ratings, demographics, and movie metadata.
The objective is to analyze user behavior and apply multiple recommendation techniques to deliver accurate and relevant movie suggestions.

The dataset used is inspired by the MovieLens-style format and includes ratings, user profiles, and movie information.

---

🎯 Business Objective

As a Data Scientist at Zee, the goal is to:

• Understand user viewing preferences
• Build scalable and accurate recommender systems
• Improve user engagement through personalization
• Compare multiple recommendation algorithms and evaluation metrics

---

📂 Dataset Description

The dataset consists of three main files:

1️⃣ Ratings File (ratings.dat)

• Format: UserID::MovieID::Rating::Timestamp
• Ratings are on a 5-star scale
• Each user has rated at least 20 movies
• Used as the core signal for preference learning

2️⃣ Users File (users.dat)

• Format: UserID::Gender::Age::Occupation::Zip-code
• Contains demographic information
• Useful for user segmentation and behavior analysis

3️⃣ Movies File (movies.dat)

• Format: MovieID::Title::Genres
• Genres are pipe (|) separated (e.g., Action|Drama|Comedy)
• Helps in content-based filtering

---

🧠 Recommendation Techniques Implemented

🔹 Collaborative Filtering

• User–User similarity
• Item–Item similarity
• Pearson Correlation
• Cosine Similarity

🔹 Content-Based Filtering

• Genre-based similarity
• User preference profiling

🔹 K-Nearest Neighbors (KNN)

• User-based KNN
• Item-based KNN
• Implemented using cosine similarity

🔹 Matrix Factorization

• SVD (Singular Value Decomposition)
• Surprise Library SVD
• CMFREC (Collective Matrix Factorization)

🔹 Embedding-Based Similarity

• User embeddings
• User–User similarity
• User–Item similarity

---

📊 Evaluation Metrics

The recommender systems are evaluated using:

• MAPE (Mean Absolute Percentage Error)
• RMSE / MSE
• NDCG (Normalized Discounted Cumulative Gain)
• MRR (Mean Reciprocal Rank)

These metrics help assess both rating prediction accuracy and ranking quality.

---

🛠️ Tech Stack & Libraries

pandas
numpy
matplotlib
seaborn
scipy
scikit-learn
surprise
cmfrec