Abstract

In this proposal, we study the problem of understandinghuman sentiments from large scale collection ofInternet images based on both image features and contextualsocial network information (such as friend comments anduser description). Despite the great strides in analyzing usersentiment based on text information, the analysis of sentimentbehind the image content has largely been ignored. Thus, we extend the significant advances in text-based sentimentprediction tasks to the higherlevel challenge of predicting theunderlying sentiments behind the images. We show that neithervisual features nor the textual features are by themselvessufficient for accurate sentiment labeling. Thus, we provide away of using both of them, and formulate sentiment predictionproblem in two scenarios: supervised and unsupervised. Wedevelop an optimization algorithm for finding a local-optimasolution under the proposed framework. With experiments ontwo large-scale datasets, we show that the proposed methodimproves significantly over existing state-of-the-art methods. In the future, we are going to incorporating more informationon the social network and explore sentiment on signed social network.

Original languageEnglish (US)
Title of host publicationProceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1584-1591
Number of pages8
ISBN (Print)9781467384926
DOIs
StatePublished - Jan 29 2016
Event15th IEEE International Conference on Data Mining Workshop, ICDMW 2015 - Atlantic City, United States
Duration: Nov 14 2015Nov 17 2015

Other

Other15th IEEE International Conference on Data Mining Workshop, ICDMW 2015
CountryUnited States
CityAtlantic City
Period11/14/1511/17/15

Fingerprint

Labeling
Experiments

Keywords

  • Computer vision
  • sentiment analysis

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Science Applications

Cite this

Wang, Y., & Li, B. (2016). Sentiment Analysis for Social Media Images. In Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015 (pp. 1584-1591). [7395864] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICDMW.2015.142

Sentiment Analysis for Social Media Images. / Wang, Yilin; Li, Baoxin.

Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015. Institute of Electrical and Electronics Engineers Inc., 2016. p. 1584-1591 7395864.

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

Wang, Y & Li, B 2016, Sentiment Analysis for Social Media Images. in Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015., 7395864, Institute of Electrical and Electronics Engineers Inc., pp. 1584-1591, 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015, Atlantic City, United States, 11/14/15. https://doi.org/10.1109/ICDMW.2015.142
Wang Y, Li B. Sentiment Analysis for Social Media Images. In Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015. Institute of Electrical and Electronics Engineers Inc. 2016. p. 1584-1591. 7395864 https://doi.org/10.1109/ICDMW.2015.142
Wang, Yilin ; Li, Baoxin. / Sentiment Analysis for Social Media Images. Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 1584-1591
@inproceedings{e6de2a13c5934b98bfdeea3ba95150c7,
title = "Sentiment Analysis for Social Media Images",
abstract = "In this proposal, we study the problem of understandinghuman sentiments from large scale collection ofInternet images based on both image features and contextualsocial network information (such as friend comments anduser description). Despite the great strides in analyzing usersentiment based on text information, the analysis of sentimentbehind the image content has largely been ignored. Thus, we extend the significant advances in text-based sentimentprediction tasks to the higherlevel challenge of predicting theunderlying sentiments behind the images. We show that neithervisual features nor the textual features are by themselvessufficient for accurate sentiment labeling. Thus, we provide away of using both of them, and formulate sentiment predictionproblem in two scenarios: supervised and unsupervised. Wedevelop an optimization algorithm for finding a local-optimasolution under the proposed framework. With experiments ontwo large-scale datasets, we show that the proposed methodimproves significantly over existing state-of-the-art methods. In the future, we are going to incorporating more informationon the social network and explore sentiment on signed social network.",
keywords = "Computer vision, sentiment analysis",
author = "Yilin Wang and Baoxin Li",
year = "2016",
month = "1",
day = "29",
doi = "10.1109/ICDMW.2015.142",
language = "English (US)",
isbn = "9781467384926",
pages = "1584--1591",
booktitle = "Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Sentiment Analysis for Social Media Images

AU - Wang, Yilin

AU - Li, Baoxin

PY - 2016/1/29

Y1 - 2016/1/29

N2 - In this proposal, we study the problem of understandinghuman sentiments from large scale collection ofInternet images based on both image features and contextualsocial network information (such as friend comments anduser description). Despite the great strides in analyzing usersentiment based on text information, the analysis of sentimentbehind the image content has largely been ignored. Thus, we extend the significant advances in text-based sentimentprediction tasks to the higherlevel challenge of predicting theunderlying sentiments behind the images. We show that neithervisual features nor the textual features are by themselvessufficient for accurate sentiment labeling. Thus, we provide away of using both of them, and formulate sentiment predictionproblem in two scenarios: supervised and unsupervised. Wedevelop an optimization algorithm for finding a local-optimasolution under the proposed framework. With experiments ontwo large-scale datasets, we show that the proposed methodimproves significantly over existing state-of-the-art methods. In the future, we are going to incorporating more informationon the social network and explore sentiment on signed social network.

AB - In this proposal, we study the problem of understandinghuman sentiments from large scale collection ofInternet images based on both image features and contextualsocial network information (such as friend comments anduser description). Despite the great strides in analyzing usersentiment based on text information, the analysis of sentimentbehind the image content has largely been ignored. Thus, we extend the significant advances in text-based sentimentprediction tasks to the higherlevel challenge of predicting theunderlying sentiments behind the images. We show that neithervisual features nor the textual features are by themselvessufficient for accurate sentiment labeling. Thus, we provide away of using both of them, and formulate sentiment predictionproblem in two scenarios: supervised and unsupervised. Wedevelop an optimization algorithm for finding a local-optimasolution under the proposed framework. With experiments ontwo large-scale datasets, we show that the proposed methodimproves significantly over existing state-of-the-art methods. In the future, we are going to incorporating more informationon the social network and explore sentiment on signed social network.

KW - Computer vision

KW - sentiment analysis

UR - http://www.scopus.com/inward/record.url?scp=84964739548&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84964739548&partnerID=8YFLogxK

U2 - 10.1109/ICDMW.2015.142

DO - 10.1109/ICDMW.2015.142

M3 - Conference contribution

SN - 9781467384926

SP - 1584

EP - 1591

BT - Proceedings - 15th IEEE International Conference on Data Mining Workshop, ICDMW 2015

PB - Institute of Electrical and Electronics Engineers Inc.

ER -