"Nobody comes here anymore, it's too crowded": predicting image popularity on Flickr

Philip J. McParlane, Yashar Moshfeghi, Joemon M. Jose

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

36 Citations (Scopus)

Abstract

Predicting popular content is a challenging problem for social media websites in order to encourage user interactions and activity. Existing works in this area, including the recommendation approach used by Flickr (called "interestingness"), consider only click through data, tags, comments and explicit user feedback in this computation. On image sharing websites, however, many images are annotated with no tags and initially, an image has no interaction data. In this case, these existing approaches fail due to lack of evidence. In this paper, we therefore focus on image popularity prediction in a cold start scenario (i.e. where there exist no, or limited, textual/interaction data), by considering an image's context, visual appearance and user context. Specifically, we predict the number of comments and views an image has based on a number of new features for this propose. Experimenting on the MIR-Flickr 1M collection, we are able to overcome the problems associated with popularity prediction in a cold start, achieving accuracy of up to 76%.
LanguageEnglish
Title of host publicationProceedings of International Conference on Multimedia Retrieval
Place of PublicationNew York, NY
Number of pages7
DOIs
Publication statusPublished - 1 Apr 2014
EventInternational Conference on Multimedia Retrieval - Glasgow, United Kingdom
Duration: 1 Apr 20144 Apr 2014

Conference

ConferenceInternational Conference on Multimedia Retrieval
CountryUnited Kingdom
CityGlasgow
Period1/04/144/04/14

Fingerprint

Websites
Feedback

Keywords

  • social media
  • Flickr
  • popularity prediction
  • user interaction

Cite this

McParlane, P. J., Moshfeghi, Y., & Jose, J. M. (2014). "Nobody comes here anymore, it's too crowded": predicting image popularity on Flickr. In Proceedings of International Conference on Multimedia Retrieval New York, NY. https://doi.org/10.1145/2578726.2578776
McParlane, Philip J. ; Moshfeghi, Yashar ; Jose, Joemon M. / "Nobody comes here anymore, it's too crowded" : predicting image popularity on Flickr. Proceedings of International Conference on Multimedia Retrieval. New York, NY, 2014.
@inproceedings{f9313b644f154e0f8b39aaa3920ed63d,
title = "{"}Nobody comes here anymore, it's too crowded{"}: predicting image popularity on Flickr",
abstract = "Predicting popular content is a challenging problem for social media websites in order to encourage user interactions and activity. Existing works in this area, including the recommendation approach used by Flickr (called {"}interestingness{"}), consider only click through data, tags, comments and explicit user feedback in this computation. On image sharing websites, however, many images are annotated with no tags and initially, an image has no interaction data. In this case, these existing approaches fail due to lack of evidence. In this paper, we therefore focus on image popularity prediction in a cold start scenario (i.e. where there exist no, or limited, textual/interaction data), by considering an image's context, visual appearance and user context. Specifically, we predict the number of comments and views an image has based on a number of new features for this propose. Experimenting on the MIR-Flickr 1M collection, we are able to overcome the problems associated with popularity prediction in a cold start, achieving accuracy of up to 76{\%}.",
keywords = "social media, Flickr, popularity prediction, user interaction",
author = "McParlane, {Philip J.} and Yashar Moshfeghi and Jose, {Joemon M.}",
year = "2014",
month = "4",
day = "1",
doi = "10.1145/2578726.2578776",
language = "English",
isbn = "9781450327824",
booktitle = "Proceedings of International Conference on Multimedia Retrieval",

}

McParlane, PJ, Moshfeghi, Y & Jose, JM 2014, "Nobody comes here anymore, it's too crowded": predicting image popularity on Flickr. in Proceedings of International Conference on Multimedia Retrieval. New York, NY, International Conference on Multimedia Retrieval, Glasgow, United Kingdom, 1/04/14. https://doi.org/10.1145/2578726.2578776

"Nobody comes here anymore, it's too crowded" : predicting image popularity on Flickr. / McParlane, Philip J.; Moshfeghi, Yashar; Jose, Joemon M.

Proceedings of International Conference on Multimedia Retrieval. New York, NY, 2014.

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

TY - GEN

T1 - "Nobody comes here anymore, it's too crowded"

T2 - predicting image popularity on Flickr

AU - McParlane, Philip J.

AU - Moshfeghi, Yashar

AU - Jose, Joemon M.

PY - 2014/4/1

Y1 - 2014/4/1

N2 - Predicting popular content is a challenging problem for social media websites in order to encourage user interactions and activity. Existing works in this area, including the recommendation approach used by Flickr (called "interestingness"), consider only click through data, tags, comments and explicit user feedback in this computation. On image sharing websites, however, many images are annotated with no tags and initially, an image has no interaction data. In this case, these existing approaches fail due to lack of evidence. In this paper, we therefore focus on image popularity prediction in a cold start scenario (i.e. where there exist no, or limited, textual/interaction data), by considering an image's context, visual appearance and user context. Specifically, we predict the number of comments and views an image has based on a number of new features for this propose. Experimenting on the MIR-Flickr 1M collection, we are able to overcome the problems associated with popularity prediction in a cold start, achieving accuracy of up to 76%.

AB - Predicting popular content is a challenging problem for social media websites in order to encourage user interactions and activity. Existing works in this area, including the recommendation approach used by Flickr (called "interestingness"), consider only click through data, tags, comments and explicit user feedback in this computation. On image sharing websites, however, many images are annotated with no tags and initially, an image has no interaction data. In this case, these existing approaches fail due to lack of evidence. In this paper, we therefore focus on image popularity prediction in a cold start scenario (i.e. where there exist no, or limited, textual/interaction data), by considering an image's context, visual appearance and user context. Specifically, we predict the number of comments and views an image has based on a number of new features for this propose. Experimenting on the MIR-Flickr 1M collection, we are able to overcome the problems associated with popularity prediction in a cold start, achieving accuracy of up to 76%.

KW - social media

KW - Flickr

KW - popularity prediction

KW - user interaction

UR - https://dl.acm.org

U2 - 10.1145/2578726.2578776

DO - 10.1145/2578726.2578776

M3 - Conference contribution book

SN - 9781450327824

BT - Proceedings of International Conference on Multimedia Retrieval

CY - New York, NY

ER -

McParlane PJ, Moshfeghi Y, Jose JM. "Nobody comes here anymore, it's too crowded": predicting image popularity on Flickr. In Proceedings of International Conference on Multimedia Retrieval. New York, NY. 2014 https://doi.org/10.1145/2578726.2578776