Probabilistic Matrix Factorization on MovieLens 100K

Overview

In this project, we use MovieLens 100K dataset. The dataset consists of 100,000 ratings from 943 users on 1,682 movies. In this project, RMSE (root-mean square error) is used as metric.

I test with 2 different data spliting: Dense and Sparse.

The data are randomly split, 80% for training/validation and 20% for testing for dense data, and for sparse data, only 20% is taken for training/validation and 20% for testing. In the training, 5-fold cross-validation is applied to choose the best hyper-parameters and evaluate the model in test set.

Run the code

Parameters

• task: ["task1" - Tune regularization parameter, "task2" - Tune number of factors, "predict" - Predict the ratings]

For dense dataset

> python main_dense.py --task=task1

For sparse dataset

> python main_sparse.py --task=task1

Note

COMP5212 - Machine Learning Programming Assignment 3 in HKUST