Personalized Movie Recommendation System
A hybrid recommendation system integrating content-based and collaborative filtering approaches
--- views
Developed a full-featured web application using Next.js, Tailwind CSS, and shadcn/ui to simulate profile selection and recommendations. Implemented data cleaning techniques and visualized key insights using Pandas, NumPy, and Matplotlib. Fine-tuned matrix factorization and LSTM models for collaborative filtering with optimized hyperparameters.
Technology
Applied machine learning · Recommendation system
Context#
A movie recommendation project exploring how content-based and collaborative approaches can work together.
Implementation#
The project record lists Python, TensorFlow, Pandas and NumPy for the recommendation work, with FastAPI and Next.js connecting it to a web interface.
Reflection#
This project provides a starting point for my interest in applied AI: connecting model behavior to a product people can interact with.
Next direction#
The research direction is explicit evaluation: document the dataset, ranking baseline, cold-start behavior and limitations before making performance claims.