Movie Recommender System

Works Based on Content Based Filtering Made this in a Jupyter Notebook using Pandas, Matplotlib, Numpy, Pickle and Streamlit

The similarity is based on cosine similarity which bests the usual Euclidean distance method since it focuses on their direction rather than magnitude, making it scale-invariant and robust to irrelevant features. The cosine similarity matrix uses a tags column which consists of cast name, director name, overview & title.

Built With

• jupyter
• numpy
• pandas
• python
• scikit-learn
• streamlit