Abstract

Sentiment analysis has been studied for decades, and it is widely used in many real applications such as media monitoring. In sentiment analysis, when addressing the problem of limited labeled data from the target domain, transfer learning, or domain adaptation, has been successfully applied, which borrows information from a relevant source domain with abundant labeled data to improve the prediction performance in the target domain. The key to transfer learning is how to model the relatedness among different domains. For sentiment analysis, a common practice is to assume similar sentiment polarity for the common keywords shared by different domains. However, existing methods largely overlooked the human factor, i.e., the users who expressed such sentiment. In this paper, we address this problem by explicitly modeling the human factor related to sentiment classification. In particular, we assume that the content generated by the same user across different domains is biased in the same way in terms of the sentiment polarity. In other words, optimistic/pessimistic users demonstrate consistent sentiment patterns, no matter what the context is. To this end, we propose a new graph-based approach named U-Cross, which models the relatedness of different domains via both the shared users and keywords. It is non-parametric and semi-supervised in nature. Furthermore, we also study the problem of shared user selection to prevent 'negative transfer'. In the experiments, we demonstrate the effectiveness of U-Cross by comparing it with existing state-of-the-art techniques on three real data sets.

Original languageEnglish (US)
Title of host publicationProceedings of the 17th SIAM International Conference on Data Mining, SDM 2017
PublisherSociety for Industrial and Applied Mathematics Publications
Pages471-479
Number of pages9
ISBN (Electronic)9781611974874
StatePublished - 2017
Event17th SIAM International Conference on Data Mining, SDM 2017 - Houston, United States
Duration: Apr 27 2017Apr 29 2017

Other

Other17th SIAM International Conference on Data Mining, SDM 2017
CountryUnited States
CityHouston
Period4/27/174/29/17

Fingerprint

Human engineering
Monitoring
Experiments

Keywords

  • Classification
  • Transfer learning
  • User modeling

ASJC Scopus subject areas

  • Software
  • Computer Science Applications

Cite this

Nelakurthi, A. R., Tong, H., Maciejewski, R., Bliss, N., & He, J. (2017). User-guided cross-domain sentiment classification. In Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017 (pp. 471-479). Society for Industrial and Applied Mathematics Publications.

User-guided cross-domain sentiment classification. / Nelakurthi, Arun Reddy; Tong, Hanghang; Maciejewski, Ross; Bliss, Nadya; He, Jingrui.

Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017. Society for Industrial and Applied Mathematics Publications, 2017. p. 471-479.

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

Nelakurthi, AR, Tong, H, Maciejewski, R, Bliss, N & He, J 2017, User-guided cross-domain sentiment classification. in Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017. Society for Industrial and Applied Mathematics Publications, pp. 471-479, 17th SIAM International Conference on Data Mining, SDM 2017, Houston, United States, 4/27/17.
Nelakurthi AR, Tong H, Maciejewski R, Bliss N, He J. User-guided cross-domain sentiment classification. In Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017. Society for Industrial and Applied Mathematics Publications. 2017. p. 471-479
Nelakurthi, Arun Reddy ; Tong, Hanghang ; Maciejewski, Ross ; Bliss, Nadya ; He, Jingrui. / User-guided cross-domain sentiment classification. Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017. Society for Industrial and Applied Mathematics Publications, 2017. pp. 471-479
@inproceedings{a7f7348fd19541dc87889a0ddf46c2c7,
title = "User-guided cross-domain sentiment classification",
abstract = "Sentiment analysis has been studied for decades, and it is widely used in many real applications such as media monitoring. In sentiment analysis, when addressing the problem of limited labeled data from the target domain, transfer learning, or domain adaptation, has been successfully applied, which borrows information from a relevant source domain with abundant labeled data to improve the prediction performance in the target domain. The key to transfer learning is how to model the relatedness among different domains. For sentiment analysis, a common practice is to assume similar sentiment polarity for the common keywords shared by different domains. However, existing methods largely overlooked the human factor, i.e., the users who expressed such sentiment. In this paper, we address this problem by explicitly modeling the human factor related to sentiment classification. In particular, we assume that the content generated by the same user across different domains is biased in the same way in terms of the sentiment polarity. In other words, optimistic/pessimistic users demonstrate consistent sentiment patterns, no matter what the context is. To this end, we propose a new graph-based approach named U-Cross, which models the relatedness of different domains via both the shared users and keywords. It is non-parametric and semi-supervised in nature. Furthermore, we also study the problem of shared user selection to prevent 'negative transfer'. In the experiments, we demonstrate the effectiveness of U-Cross by comparing it with existing state-of-the-art techniques on three real data sets.",
keywords = "Classification, Transfer learning, User modeling",
author = "Nelakurthi, {Arun Reddy} and Hanghang Tong and Ross Maciejewski and Nadya Bliss and Jingrui He",
year = "2017",
language = "English (US)",
pages = "471--479",
booktitle = "Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017",
publisher = "Society for Industrial and Applied Mathematics Publications",
address = "United States",

}

TY - GEN

T1 - User-guided cross-domain sentiment classification

AU - Nelakurthi, Arun Reddy

AU - Tong, Hanghang

AU - Maciejewski, Ross

AU - Bliss, Nadya

AU - He, Jingrui

PY - 2017

Y1 - 2017

N2 - Sentiment analysis has been studied for decades, and it is widely used in many real applications such as media monitoring. In sentiment analysis, when addressing the problem of limited labeled data from the target domain, transfer learning, or domain adaptation, has been successfully applied, which borrows information from a relevant source domain with abundant labeled data to improve the prediction performance in the target domain. The key to transfer learning is how to model the relatedness among different domains. For sentiment analysis, a common practice is to assume similar sentiment polarity for the common keywords shared by different domains. However, existing methods largely overlooked the human factor, i.e., the users who expressed such sentiment. In this paper, we address this problem by explicitly modeling the human factor related to sentiment classification. In particular, we assume that the content generated by the same user across different domains is biased in the same way in terms of the sentiment polarity. In other words, optimistic/pessimistic users demonstrate consistent sentiment patterns, no matter what the context is. To this end, we propose a new graph-based approach named U-Cross, which models the relatedness of different domains via both the shared users and keywords. It is non-parametric and semi-supervised in nature. Furthermore, we also study the problem of shared user selection to prevent 'negative transfer'. In the experiments, we demonstrate the effectiveness of U-Cross by comparing it with existing state-of-the-art techniques on three real data sets.

AB - Sentiment analysis has been studied for decades, and it is widely used in many real applications such as media monitoring. In sentiment analysis, when addressing the problem of limited labeled data from the target domain, transfer learning, or domain adaptation, has been successfully applied, which borrows information from a relevant source domain with abundant labeled data to improve the prediction performance in the target domain. The key to transfer learning is how to model the relatedness among different domains. For sentiment analysis, a common practice is to assume similar sentiment polarity for the common keywords shared by different domains. However, existing methods largely overlooked the human factor, i.e., the users who expressed such sentiment. In this paper, we address this problem by explicitly modeling the human factor related to sentiment classification. In particular, we assume that the content generated by the same user across different domains is biased in the same way in terms of the sentiment polarity. In other words, optimistic/pessimistic users demonstrate consistent sentiment patterns, no matter what the context is. To this end, we propose a new graph-based approach named U-Cross, which models the relatedness of different domains via both the shared users and keywords. It is non-parametric and semi-supervised in nature. Furthermore, we also study the problem of shared user selection to prevent 'negative transfer'. In the experiments, we demonstrate the effectiveness of U-Cross by comparing it with existing state-of-the-art techniques on three real data sets.

KW - Classification

KW - Transfer learning

KW - User modeling

UR - http://www.scopus.com/inward/record.url?scp=85027876536&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027876536&partnerID=8YFLogxK

M3 - Conference contribution

SP - 471

EP - 479

BT - Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017

PB - Society for Industrial and Applied Mathematics Publications

ER -