In this project, we will show you how to do ranking prediction from the MovieLens 1m dataset. First, we will prepare the dataset. Then, we will train the FM algorithm, which eventually predicts the rankings and ratings for movies. The project code has the following structure:
Movie rating and ranking prediction project structure
In summary, the project has the following structure:
- EDA: This package is used to do an exploratory analysis of the MovieLens 1M dataset.
- Tools, FMCore, and DataUtils: These are the core FM libraries. For the purpose of this probject, I used (but extended) the RankSys library (see the GitHub repository at https://github.com/RankSys/RankSys).
- Preprocessing: This package is used to convert the MovieLens 1M dataset into LibFM format.
- Prediction: This package is used for the movie rating and ranking prediction...
