Inspired by the Netflix Recommendation Challenge, where a $1,000,000 prize was offered to the first team to achieve an improve of 10% or better over Netflix's then deployed movie recommendation algorithm, this project covers the development of a matrix factorization machine learning model with an RMSE score of 0.781 (approximately 9.6% better than Netflix's model). Created independently in Summer 2022, the project utilized R, RStudio, machine learning, data visualization, and data cleaning techniques to both outperform Netflix's algorithm and make explicit the power of matrix factorization in predictive models.
This project was completed in July 2022 and no further work is expected at this time.
I make no claim to the data used. The dataset is available at: https://grouplens.org/datasets/movielens/10m/.