HarvardX DataScience Professional Certificate - Final Capstone Project
In this project, we created a movie recommendation system using a large subset of the MovieLens dataset. Exploration of the edx dataset showed strong evidence of a genre effect and if used to augment the Regularized Movie + User Effect Model could yield improvements in the accuracy of the model. This hypothesis was validated when the final model, with the addition of an regularized predicted genre effect "bias" term, yielded a residual mean squared error (RMSE) of 0.8575 on the validation dataset.
MovieLensProj-code.R - This R file needs to be ran first to create and store the datafiles that feeds the .rmd file below.
MovieLensProj.rmd - This .rmd file creates the fully reproducible final project .pdf.
MovieLensProj.pdf - Final Project .pdf file.