This project implements a content-based movie recommendation system utilizing the TMDB 5000 dataset from Kaggle. The system analyzes various movie attributes to generate personalized recommendations based on user input.
Source: TMDB 5000 Dataset (https://www.kaggle.com/datasets/tmdb/tmdb-movie-metadata)
Version 1: Stemming + Bag of Words + Similarity Search Version 2: Lemmatization + TFIDF + Similarity Search
- Data Preprocessing: Clean and prepare the dataset for analysis.
- Exploratory Data Analysis (EDA): Analyze the dataset to understand its structure and key features.
- Feature Engineering: Extract meaningful features from the dataset to enhance recommendation accuracy.
- Tag Creation: Generate tags based on multiple columns including Genre, Overview, Keywords, Cast, and Crew.
- Text Processing: Apply stemming and lemmatization techniques, and remove stop words to refine the tags for better similarity matching.
- Cosine similarity is applied to find the similar movies