A Hybrid Feature Combination Method that Improves Recommendations

Gharbi Alshammari, Stylianos Kapetanakis, Abduallah Alshammari, Nikolaos Polatidis, Miltos Petridis

Research output: Chapter in Book/Conference proceeding with ISSN or ISBNConference contribution with ISSN or ISBNResearchpeer-review

Abstract

Recommender systems help users find relevant items efficiently based on their interests and historical interactions. They can also be beneficial to businesses by promoting the sale of products. Recommender systems can be modelled by applying different approaches, including collaborative filtering (CF), demographic filtering (DF), content-based filtering (CBF) and knowledge-based filtering (KBF). However, large amounts of data can produce recommendations that are limited in accuracy because of diversity and sparsity issues. In this paper, we propose a novel hybrid approach that combines user-user CF with the attributes of DF to indicate the nearest users, and compare the Random Forest classifier against the kNN classifier, developed through an investigation of ways to reduce the errors in rating predictions based on users past interactions. Our combined method leads to improved prediction accuracy in two different classification algorithms. The main goal of this paper is to identify the impact of DF on CF and compare the two classifiers. We apply a feature combination hybrid method that can improve prediction accuracy and achieve lower mean absolute error values compared with the results of CF or DF alone. To test our approach, we ran an offline evaluation using the 1 M MovieLens data set.
Original languageEnglish
Title of host publicationInternational Conference on Computational Collective Intelligence
PublisherSpringer
Pages209-218
Number of pages10
ISBN (Print)9783319984421
DOIs
Publication statusPublished - 8 Aug 2018
EventInternational Conference on Computational Collective Intelligence - Bristol, United Kingdom
Duration: 5 Sep 20187 Sep 2018
Conference number: 10
http://iccci2018.org/
http://www.iccci2018.org

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
ISSN (Print)0302-9743

Conference

ConferenceInternational Conference on Computational Collective Intelligence
Abbreviated titleICCCI
CountryUnited Kingdom
CityBristol
Period5/09/187/09/18
Internet address

Fingerprint

Collaborative filtering
Classifiers
Recommender systems
Sales
Industry

Cite this

Alshammari, G., Kapetanakis, S., Alshammari, A., Polatidis, N., & Petridis, M. (2018). A Hybrid Feature Combination Method that Improves Recommendations. In International Conference on Computational Collective Intelligence (pp. 209-218). (Lecture Notes in Computer Science). Springer. https://doi.org/10.1007/978-3-319-98443-8_19
Alshammari, Gharbi ; Kapetanakis, Stylianos ; Alshammari, Abduallah ; Polatidis, Nikolaos ; Petridis, Miltos. / A Hybrid Feature Combination Method that Improves Recommendations. International Conference on Computational Collective Intelligence. Springer, 2018. pp. 209-218 (Lecture Notes in Computer Science).
@inproceedings{75c520b99f3441d88daa244e952c9c7d,
title = "A Hybrid Feature Combination Method that Improves Recommendations",
abstract = "Recommender systems help users find relevant items efficiently based on their interests and historical interactions. They can also be beneficial to businesses by promoting the sale of products. Recommender systems can be modelled by applying different approaches, including collaborative filtering (CF), demographic filtering (DF), content-based filtering (CBF) and knowledge-based filtering (KBF). However, large amounts of data can produce recommendations that are limited in accuracy because of diversity and sparsity issues. In this paper, we propose a novel hybrid approach that combines user-user CF with the attributes of DF to indicate the nearest users, and compare the Random Forest classifier against the kNN classifier, developed through an investigation of ways to reduce the errors in rating predictions based on users past interactions. Our combined method leads to improved prediction accuracy in two different classification algorithms. The main goal of this paper is to identify the impact of DF on CF and compare the two classifiers. We apply a feature combination hybrid method that can improve prediction accuracy and achieve lower mean absolute error values compared with the results of CF or DF alone. To test our approach, we ran an offline evaluation using the 1 M MovieLens data set.",
author = "Gharbi Alshammari and Stylianos Kapetanakis and Abduallah Alshammari and Nikolaos Polatidis and Miltos Petridis",
year = "2018",
month = "8",
day = "8",
doi = "10.1007/978-3-319-98443-8_19",
language = "English",
isbn = "9783319984421",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "209--218",
booktitle = "International Conference on Computational Collective Intelligence",

}

Alshammari, G, Kapetanakis, S, Alshammari, A, Polatidis, N & Petridis, M 2018, A Hybrid Feature Combination Method that Improves Recommendations. in International Conference on Computational Collective Intelligence. Lecture Notes in Computer Science, Springer, pp. 209-218, International Conference on Computational Collective Intelligence, Bristol, United Kingdom, 5/09/18. https://doi.org/10.1007/978-3-319-98443-8_19

A Hybrid Feature Combination Method that Improves Recommendations. / Alshammari, Gharbi; Kapetanakis, Stylianos; Alshammari, Abduallah ; Polatidis, Nikolaos; Petridis, Miltos.

International Conference on Computational Collective Intelligence. Springer, 2018. p. 209-218 (Lecture Notes in Computer Science).

Research output: Chapter in Book/Conference proceeding with ISSN or ISBNConference contribution with ISSN or ISBNResearchpeer-review

TY - GEN

T1 - A Hybrid Feature Combination Method that Improves Recommendations

AU - Alshammari, Gharbi

AU - Kapetanakis, Stylianos

AU - Alshammari, Abduallah

AU - Polatidis, Nikolaos

AU - Petridis, Miltos

PY - 2018/8/8

Y1 - 2018/8/8

N2 - Recommender systems help users find relevant items efficiently based on their interests and historical interactions. They can also be beneficial to businesses by promoting the sale of products. Recommender systems can be modelled by applying different approaches, including collaborative filtering (CF), demographic filtering (DF), content-based filtering (CBF) and knowledge-based filtering (KBF). However, large amounts of data can produce recommendations that are limited in accuracy because of diversity and sparsity issues. In this paper, we propose a novel hybrid approach that combines user-user CF with the attributes of DF to indicate the nearest users, and compare the Random Forest classifier against the kNN classifier, developed through an investigation of ways to reduce the errors in rating predictions based on users past interactions. Our combined method leads to improved prediction accuracy in two different classification algorithms. The main goal of this paper is to identify the impact of DF on CF and compare the two classifiers. We apply a feature combination hybrid method that can improve prediction accuracy and achieve lower mean absolute error values compared with the results of CF or DF alone. To test our approach, we ran an offline evaluation using the 1 M MovieLens data set.

AB - Recommender systems help users find relevant items efficiently based on their interests and historical interactions. They can also be beneficial to businesses by promoting the sale of products. Recommender systems can be modelled by applying different approaches, including collaborative filtering (CF), demographic filtering (DF), content-based filtering (CBF) and knowledge-based filtering (KBF). However, large amounts of data can produce recommendations that are limited in accuracy because of diversity and sparsity issues. In this paper, we propose a novel hybrid approach that combines user-user CF with the attributes of DF to indicate the nearest users, and compare the Random Forest classifier against the kNN classifier, developed through an investigation of ways to reduce the errors in rating predictions based on users past interactions. Our combined method leads to improved prediction accuracy in two different classification algorithms. The main goal of this paper is to identify the impact of DF on CF and compare the two classifiers. We apply a feature combination hybrid method that can improve prediction accuracy and achieve lower mean absolute error values compared with the results of CF or DF alone. To test our approach, we ran an offline evaluation using the 1 M MovieLens data set.

U2 - 10.1007/978-3-319-98443-8_19

DO - 10.1007/978-3-319-98443-8_19

M3 - Conference contribution with ISSN or ISBN

SN - 9783319984421

T3 - Lecture Notes in Computer Science

SP - 209

EP - 218

BT - International Conference on Computational Collective Intelligence

PB - Springer

ER -

Alshammari G, Kapetanakis S, Alshammari A, Polatidis N, Petridis M. A Hybrid Feature Combination Method that Improves Recommendations. In International Conference on Computational Collective Intelligence. Springer. 2018. p. 209-218. (Lecture Notes in Computer Science). https://doi.org/10.1007/978-3-319-98443-8_19