Ranking item features by mining online user-item interactions

Sofiane Abbar, Habibur Rahman, Saravanan Thirumuruganathan, Carlos Castillo, Gautam Das

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We assume a database of items in which each item is described by a set of attributes, some of which could be multi-valued. We refer to each of the distinct attribute values as a feature. We also assume that we have information about the interactions (such as visits or likes) between a set of users and those items. In our paper, we would like to rank the features of an item using user-item interactions. For instance, if the items are movies, features could be actors, directors or genres, and user-item interaction could be user liking the movie. These information could be used to identify the most important actors for each movie. While users are drawn to an item due to a subset of its features, a user-item interaction only provides an expression of user preference over the entire item, and not its component features. We design algorithms to rank the features of an item depending on whether interaction information is available at aggregated or individual level granularity and extend them to rank composite features (set of features). Our algorithms are based on constrained least squares, network flow and non-trivial adaptations to non-negative matrix factorization. We evaluate our algorithms using both real-world and synthetic datasets.

Original languageEnglish
Title of host publicationProceedings - International Conference on Data Engineering
PublisherIEEE Computer Society
Pages460-471
Number of pages12
ISBN (Print)9781479925544
DOIs
Publication statusPublished - 1 Jan 2014
Event30th IEEE International Conference on Data Engineering, ICDE 2014 - Chicago, IL, United States
Duration: 31 Mar 20144 Apr 2014

Other

Other30th IEEE International Conference on Data Engineering, ICDE 2014
CountryUnited States
CityChicago, IL
Period31/3/144/4/14

Fingerprint

Factorization
Composite materials

ASJC Scopus subject areas

  • Information Systems
  • Signal Processing
  • Software

Cite this

Abbar, S., Rahman, H., Thirumuruganathan, S., Castillo, C., & Das, G. (2014). Ranking item features by mining online user-item interactions. In Proceedings - International Conference on Data Engineering (pp. 460-471). [6816673] IEEE Computer Society. https://doi.org/10.1109/ICDE.2014.6816673

Ranking item features by mining online user-item interactions. / Abbar, Sofiane; Rahman, Habibur; Thirumuruganathan, Saravanan; Castillo, Carlos; Das, Gautam.

Proceedings - International Conference on Data Engineering. IEEE Computer Society, 2014. p. 460-471 6816673.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abbar, S, Rahman, H, Thirumuruganathan, S, Castillo, C & Das, G 2014, Ranking item features by mining online user-item interactions. in Proceedings - International Conference on Data Engineering., 6816673, IEEE Computer Society, pp. 460-471, 30th IEEE International Conference on Data Engineering, ICDE 2014, Chicago, IL, United States, 31/3/14. https://doi.org/10.1109/ICDE.2014.6816673
Abbar S, Rahman H, Thirumuruganathan S, Castillo C, Das G. Ranking item features by mining online user-item interactions. In Proceedings - International Conference on Data Engineering. IEEE Computer Society. 2014. p. 460-471. 6816673 https://doi.org/10.1109/ICDE.2014.6816673
Abbar, Sofiane ; Rahman, Habibur ; Thirumuruganathan, Saravanan ; Castillo, Carlos ; Das, Gautam. / Ranking item features by mining online user-item interactions. Proceedings - International Conference on Data Engineering. IEEE Computer Society, 2014. pp. 460-471
@inproceedings{99ed14fcd7554277b4c13e37486b7330,
title = "Ranking item features by mining online user-item interactions",
abstract = "We assume a database of items in which each item is described by a set of attributes, some of which could be multi-valued. We refer to each of the distinct attribute values as a feature. We also assume that we have information about the interactions (such as visits or likes) between a set of users and those items. In our paper, we would like to rank the features of an item using user-item interactions. For instance, if the items are movies, features could be actors, directors or genres, and user-item interaction could be user liking the movie. These information could be used to identify the most important actors for each movie. While users are drawn to an item due to a subset of its features, a user-item interaction only provides an expression of user preference over the entire item, and not its component features. We design algorithms to rank the features of an item depending on whether interaction information is available at aggregated or individual level granularity and extend them to rank composite features (set of features). Our algorithms are based on constrained least squares, network flow and non-trivial adaptations to non-negative matrix factorization. We evaluate our algorithms using both real-world and synthetic datasets.",
author = "Sofiane Abbar and Habibur Rahman and Saravanan Thirumuruganathan and Carlos Castillo and Gautam Das",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/ICDE.2014.6816673",
language = "English",
isbn = "9781479925544",
pages = "460--471",
booktitle = "Proceedings - International Conference on Data Engineering",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Ranking item features by mining online user-item interactions

AU - Abbar, Sofiane

AU - Rahman, Habibur

AU - Thirumuruganathan, Saravanan

AU - Castillo, Carlos

AU - Das, Gautam

PY - 2014/1/1

Y1 - 2014/1/1

N2 - We assume a database of items in which each item is described by a set of attributes, some of which could be multi-valued. We refer to each of the distinct attribute values as a feature. We also assume that we have information about the interactions (such as visits or likes) between a set of users and those items. In our paper, we would like to rank the features of an item using user-item interactions. For instance, if the items are movies, features could be actors, directors or genres, and user-item interaction could be user liking the movie. These information could be used to identify the most important actors for each movie. While users are drawn to an item due to a subset of its features, a user-item interaction only provides an expression of user preference over the entire item, and not its component features. We design algorithms to rank the features of an item depending on whether interaction information is available at aggregated or individual level granularity and extend them to rank composite features (set of features). Our algorithms are based on constrained least squares, network flow and non-trivial adaptations to non-negative matrix factorization. We evaluate our algorithms using both real-world and synthetic datasets.

AB - We assume a database of items in which each item is described by a set of attributes, some of which could be multi-valued. We refer to each of the distinct attribute values as a feature. We also assume that we have information about the interactions (such as visits or likes) between a set of users and those items. In our paper, we would like to rank the features of an item using user-item interactions. For instance, if the items are movies, features could be actors, directors or genres, and user-item interaction could be user liking the movie. These information could be used to identify the most important actors for each movie. While users are drawn to an item due to a subset of its features, a user-item interaction only provides an expression of user preference over the entire item, and not its component features. We design algorithms to rank the features of an item depending on whether interaction information is available at aggregated or individual level granularity and extend them to rank composite features (set of features). Our algorithms are based on constrained least squares, network flow and non-trivial adaptations to non-negative matrix factorization. We evaluate our algorithms using both real-world and synthetic datasets.

UR - http://www.scopus.com/inward/record.url?scp=84901751700&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84901751700&partnerID=8YFLogxK

U2 - 10.1109/ICDE.2014.6816673

DO - 10.1109/ICDE.2014.6816673

M3 - Conference contribution

AN - SCOPUS:84901751700

SN - 9781479925544

SP - 460

EP - 471

BT - Proceedings - International Conference on Data Engineering

PB - IEEE Computer Society

ER -