/MusicRecommenderEvaluator

Evaluation Framework for the Music Recommender System presented at the ICMR 2017

Primary LanguageRGNU Lesser General Public License v3.0LGPL-3.0

MusicRecommenderEvaluator

Evaluation Framework for the Music Recommender System presented at the ICMR 2017.

Prerequisites

In case you want to compute the predicted ratings of the items using the original libFM implementation provided by S. Rendle, please install libFM and set the lfmpath variable accordingly. Alternatively, you can use the R implementation FactoRizationMachines by J. Knoll.

Input data

The input read by the function loadData must be a CSV file formatted according to the following scheme:

<user,artist,track,cluster>

The IDs can be arbitrary strings or integers, uniquely identifying the users, artists, tracks and clusters.

Start Evaluation

To start the evaluation, please adapt and run the main_icmr17.R. The R script will automatically install or update all required R packages.