A novel recommendation method based on general matrix factorization and artificial neural networks

Stelios Kapetanakis, Nikolaos Polatidis, Gharbi Alshammari, Miltiadis Petridis

Research output: Contribution to journalArticle

Abstract

Collaborative filtering is a successful approach in relevant item or service recommendation provision to users in rich, online domains. This approach has been widely applied in commercial environments with success, especially in online marketing, similar product suggestion and selection and tailor-made consumer suggestions. However, regardless of its market penetration, there are still considerable limitations in terms of accuracy in the proposed recommendations stemming from the high-frequency low-relevance user-item bias, data specificities and individual user patterns and needs that may be hidden in data. We propose a novel recommendation approach that improves accuracy and requires significantly less maintenance compared to traditional collaborative filtering. For the experimental evaluation, we use two real data sets and well-known metrics with the results validating our method. Our proposed method outperforms all the alternative recommendation methods for each of the two data sets and metrics and seems holistically effective against alternatives since it requires fewer settings to be considered without affecting the output.

Original languageEnglish
Pages (from-to)1-8
JournalNeural Computing and Applications
DOIs
Publication statusPublished - 26 Oct 2019

Fingerprint

Collaborative filtering
Factorization
Neural networks
Marketing

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/s00521-019-04534-w

Keywords

  • Recommender Systems
  • Collaborative Filtering
  • Matrix Factorization
  • Neural Networks
  • Collaborative filtering
  • Neural networks
  • Matrix factorization
  • Recommender systems

Cite this

@article{29fbdd2186594adeb7359b6eb928cf64,
title = "A novel recommendation method based on general matrix factorization and artificial neural networks",
abstract = "Collaborative filtering is a successful approach in relevant item or service recommendation provision to users in rich, online domains. This approach has been widely applied in commercial environments with success, especially in online marketing, similar product suggestion and selection and tailor-made consumer suggestions. However, regardless of its market penetration, there are still considerable limitations in terms of accuracy in the proposed recommendations stemming from the high-frequency low-relevance user-item bias, data specificities and individual user patterns and needs that may be hidden in data. We propose a novel recommendation approach that improves accuracy and requires significantly less maintenance compared to traditional collaborative filtering. For the experimental evaluation, we use two real data sets and well-known metrics with the results validating our method. Our proposed method outperforms all the alternative recommendation methods for each of the two data sets and metrics and seems holistically effective against alternatives since it requires fewer settings to be considered without affecting the output.",
keywords = "Recommender Systems, Collaborative Filtering, Matrix Factorization, Neural Networks, Collaborative filtering, Neural networks, Matrix factorization, Recommender systems",
author = "Stelios Kapetanakis and Nikolaos Polatidis and Gharbi Alshammari and Miltiadis Petridis",
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/s00521-019-04534-w",
year = "2019",
month = "10",
day = "26",
doi = "10.1007/s00521-019-04534-w",
language = "English",
pages = "1--8",
journal = "Neural Computing and Applications",
issn = "1433-3058",
publisher = "Springer",

}

A novel recommendation method based on general matrix factorization and artificial neural networks. / Kapetanakis, Stelios; Polatidis, Nikolaos; Alshammari, Gharbi; Petridis, Miltiadis.

In: Neural Computing and Applications, 26.10.2019, p. 1-8.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A novel recommendation method based on general matrix factorization and artificial neural networks

AU - Kapetanakis, Stelios

AU - Polatidis, Nikolaos

AU - Alshammari, Gharbi

AU - Petridis, Miltiadis

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/s00521-019-04534-w

PY - 2019/10/26

Y1 - 2019/10/26

N2 - Collaborative filtering is a successful approach in relevant item or service recommendation provision to users in rich, online domains. This approach has been widely applied in commercial environments with success, especially in online marketing, similar product suggestion and selection and tailor-made consumer suggestions. However, regardless of its market penetration, there are still considerable limitations in terms of accuracy in the proposed recommendations stemming from the high-frequency low-relevance user-item bias, data specificities and individual user patterns and needs that may be hidden in data. We propose a novel recommendation approach that improves accuracy and requires significantly less maintenance compared to traditional collaborative filtering. For the experimental evaluation, we use two real data sets and well-known metrics with the results validating our method. Our proposed method outperforms all the alternative recommendation methods for each of the two data sets and metrics and seems holistically effective against alternatives since it requires fewer settings to be considered without affecting the output.

AB - Collaborative filtering is a successful approach in relevant item or service recommendation provision to users in rich, online domains. This approach has been widely applied in commercial environments with success, especially in online marketing, similar product suggestion and selection and tailor-made consumer suggestions. However, regardless of its market penetration, there are still considerable limitations in terms of accuracy in the proposed recommendations stemming from the high-frequency low-relevance user-item bias, data specificities and individual user patterns and needs that may be hidden in data. We propose a novel recommendation approach that improves accuracy and requires significantly less maintenance compared to traditional collaborative filtering. For the experimental evaluation, we use two real data sets and well-known metrics with the results validating our method. Our proposed method outperforms all the alternative recommendation methods for each of the two data sets and metrics and seems holistically effective against alternatives since it requires fewer settings to be considered without affecting the output.

KW - Recommender Systems

KW - Collaborative Filtering

KW - Matrix Factorization

KW - Neural Networks

KW - Collaborative filtering

KW - Neural networks

KW - Matrix factorization

KW - Recommender systems

UR - http://www.scopus.com/inward/record.url?scp=85074618164&partnerID=8YFLogxK

U2 - 10.1007/s00521-019-04534-w

DO - 10.1007/s00521-019-04534-w

M3 - Article

SP - 1

EP - 8

JO - Neural Computing and Applications

JF - Neural Computing and Applications

SN - 1433-3058

ER -