socialcomplab/recsys21-relistening-actr

Code for the paper: Predicting Music Relistening Behavior Using the ACT-R Framework

Predicting Music Relistening Behavior Using the ACT-R Framework

This repository contains the code for the paper: Predicting Music Relistening Behavior Using the ACT-R Framework. (Link to paper or Link to preprint)

The paper deals with modeling relistening behavior using a cogntive framework of human memory.

TLDR: we demonstrate the utility of modeling the relistening behavior using the ACT-R framework.

Structure

The repo consists of 5 stages:

• First everything regarding the dataset itself including preparation, user selection, and statistics
• Analyzing the relistening behavior
• The evaluation protocol
• The performance comparison
• Considering combinations including parameter estimation, evaluation, and reporting the performance

In addition we provide the code for the data splitting and the algorithms.

Cite

Bibtex for Paper:

@inproceedings{reiter2021predicting,
  title={Predicting Music Relistening Behavior Using the ACT-R Framework},
  author={Reiter-Haas, Markus and Parada-Cabaleiro, Emilia and Schedl, Markus and Motamedi, Elham and Tkalcic, Marko and Lex, Elisabeth},
  booktitle={Fifteenth ACM Conference on Recommender Systems},
  pages={702--707},
  year={2021}
}