An explanation-based approach for experiment reproducibility in recommender systems

Nikolaos Polatidis, Antonios Papaleonidas, Elias Pimenidis, Lazaros Iliadis

Research output: Contribution to journalArticleResearchpeer-review

Abstract

The offline evaluation of recommender systems is typically based on accuracy metrics such as the Mean Absolute Error (MAE) and the Root Mean Squared Error (RMSE) for error rating prediction and Precision and Recall for measuring the quality of the top-N recommendations. However, it is difficult to reproduce the results since there are various libraries that can be used for running experiments and also within the same library there are many different settings that if not taken into consideration when replicating the results might vary. In this paper, we show that within the use of the same library an explanation-based approach can be used to assist in the reproducibility of experiments. Our proposed approach has been experimentally evaluated using a wide range of recommendation algorithms ranging from collaborative filtering to complicated fuzzy recommendation approaches that can solve the filter bubble problem, a real dataset and the results show that it is both practical and effective.
Original languageEnglish
Pages (from-to)1-8
JournalNeural Computing and Applications
DOIs
Publication statusPublished - 27 May 2019

Fingerprint

Recommender systems
Collaborative filtering
Experiments

Bibliographical note

This is a post-peer-review, pre-copyedit version of an article published in Neural Computing and Applications. The final authenticated version is available online at: http://dx.doi.org/10.1007/s00521-019-04274-x

Cite this

Polatidis, Nikolaos ; Papaleonidas, Antonios ; Pimenidis, Elias ; Iliadis, Lazaros. / An explanation-based approach for experiment reproducibility in recommender systems. In: Neural Computing and Applications. 2019 ; pp. 1-8.
@article{71e18a965f6a4c8dad1166dbeab143d5,
title = "An explanation-based approach for experiment reproducibility in recommender systems",
abstract = "The offline evaluation of recommender systems is typically based on accuracy metrics such as the Mean Absolute Error (MAE) and the Root Mean Squared Error (RMSE) for error rating prediction and Precision and Recall for measuring the quality of the top-N recommendations. However, it is difficult to reproduce the results since there are various libraries that can be used for running experiments and also within the same library there are many different settings that if not taken into consideration when replicating the results might vary. In this paper, we show that within the use of the same library an explanation-based approach can be used to assist in the reproducibility of experiments. Our proposed approach has been experimentally evaluated using a wide range of recommendation algorithms ranging from collaborative filtering to complicated fuzzy recommendation approaches that can solve the filter bubble problem, a real dataset and the results show that it is both practical and effective.",
author = "Nikolaos Polatidis and Antonios Papaleonidas and Elias Pimenidis and Lazaros Iliadis",
note = "This is a post-peer-review, pre-copyedit version of an article published in Neural Computing and Applications. The final authenticated version is available online at: http://dx.doi.org/10.1007/s00521-019-04274-x",
year = "2019",
month = "5",
day = "27",
doi = "10.1007/s00521-019-04274-x",
language = "English",
pages = "1--8",
journal = "Neural Computing and Applications",
issn = "1433-3058",
publisher = "Springer",

}

An explanation-based approach for experiment reproducibility in recommender systems. / Polatidis, Nikolaos; Papaleonidas, Antonios; Pimenidis, Elias; Iliadis, Lazaros.

In: Neural Computing and Applications, 27.05.2019, p. 1-8.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - An explanation-based approach for experiment reproducibility in recommender systems

AU - Polatidis, Nikolaos

AU - Papaleonidas, Antonios

AU - Pimenidis, Elias

AU - Iliadis, Lazaros

N1 - This is a post-peer-review, pre-copyedit version of an article published in Neural Computing and Applications. The final authenticated version is available online at: http://dx.doi.org/10.1007/s00521-019-04274-x

PY - 2019/5/27

Y1 - 2019/5/27

N2 - The offline evaluation of recommender systems is typically based on accuracy metrics such as the Mean Absolute Error (MAE) and the Root Mean Squared Error (RMSE) for error rating prediction and Precision and Recall for measuring the quality of the top-N recommendations. However, it is difficult to reproduce the results since there are various libraries that can be used for running experiments and also within the same library there are many different settings that if not taken into consideration when replicating the results might vary. In this paper, we show that within the use of the same library an explanation-based approach can be used to assist in the reproducibility of experiments. Our proposed approach has been experimentally evaluated using a wide range of recommendation algorithms ranging from collaborative filtering to complicated fuzzy recommendation approaches that can solve the filter bubble problem, a real dataset and the results show that it is both practical and effective.

AB - The offline evaluation of recommender systems is typically based on accuracy metrics such as the Mean Absolute Error (MAE) and the Root Mean Squared Error (RMSE) for error rating prediction and Precision and Recall for measuring the quality of the top-N recommendations. However, it is difficult to reproduce the results since there are various libraries that can be used for running experiments and also within the same library there are many different settings that if not taken into consideration when replicating the results might vary. In this paper, we show that within the use of the same library an explanation-based approach can be used to assist in the reproducibility of experiments. Our proposed approach has been experimentally evaluated using a wide range of recommendation algorithms ranging from collaborative filtering to complicated fuzzy recommendation approaches that can solve the filter bubble problem, a real dataset and the results show that it is both practical and effective.

U2 - 10.1007/s00521-019-04274-x

DO - 10.1007/s00521-019-04274-x

M3 - Article

SP - 1

EP - 8

JO - Neural Computing and Applications

T2 - Neural Computing and Applications

JF - Neural Computing and Applications

SN - 1433-3058

ER -