Details
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 33rd Annual ACM Symposium on Applied Computing, SAC 2018 |
| Pages | 525-528 |
| Number of pages | 4 |
| ISBN (electronic) | 9781450351911 |
| Publication status | Published - 9 Apr 2018 |
| Event | 33rd Annual ACM Symposium on Applied Computing, SAC 2018 - Pau, France Duration: 9 Apr 2018 → 13 Apr 2018 |
Abstract
Opinion stream mining algorithms learn and adapt a polarity model as new opinionated texts arrive. Text understanding is computationally expensive though, and sensitive to the emergence of new words. In this work, we study polarity prediction for opinions on given entities and investigate how prediction quality is affected when we ignore the text of past opinions but exploit the entity-opinion link and the past polarity scores on it. We model each entity as a trajectory of polarity scores and propose learning algorithms that exploit these trajectories for polarity prediction. We study the performance of our approach on the Tools & Home Improvement products of the Amazon Reviews Dataset1.
Keywords
- Document polarity, Prediction, Stream mining, Trajectory
ASJC Scopus subject areas
- Computer Science(all)
- Software
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Proceedings of the 33rd Annual ACM Symposium on Applied Computing, SAC 2018. 2018. p. 525-528.
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Predicting Polarities of Entity-Centered Documents without Reading their Contents
AU - Beyer, Christian
AU - Niemann, Uli
AU - Unnikrishnan, Vishnu
AU - Ntoutsi, Eirini
AU - Spiliopoulou, Myra
N1 - Funding information: This work is partially funded by the German Research Foundation, project OSCAR “Opinion Stream Classification with Ensembles and Active Learners”. Additionally, the first author is also partially funded by a PhD grant from the federal state of Saxony-Anhalt.
PY - 2018/4/9
Y1 - 2018/4/9
N2 - Opinion stream mining algorithms learn and adapt a polarity model as new opinionated texts arrive. Text understanding is computationally expensive though, and sensitive to the emergence of new words. In this work, we study polarity prediction for opinions on given entities and investigate how prediction quality is affected when we ignore the text of past opinions but exploit the entity-opinion link and the past polarity scores on it. We model each entity as a trajectory of polarity scores and propose learning algorithms that exploit these trajectories for polarity prediction. We study the performance of our approach on the Tools & Home Improvement products of the Amazon Reviews Dataset1.
AB - Opinion stream mining algorithms learn and adapt a polarity model as new opinionated texts arrive. Text understanding is computationally expensive though, and sensitive to the emergence of new words. In this work, we study polarity prediction for opinions on given entities and investigate how prediction quality is affected when we ignore the text of past opinions but exploit the entity-opinion link and the past polarity scores on it. We model each entity as a trajectory of polarity scores and propose learning algorithms that exploit these trajectories for polarity prediction. We study the performance of our approach on the Tools & Home Improvement products of the Amazon Reviews Dataset1.
KW - Document polarity
KW - Prediction
KW - Stream mining
KW - Trajectory
UR - http://www.scopus.com/inward/record.url?scp=85050558198&partnerID=8YFLogxK
U2 - 10.1145/3167132.3172870
DO - 10.1145/3167132.3172870
M3 - Conference contribution
AN - SCOPUS:85050558198
SP - 525
EP - 528
BT - Proceedings of the 33rd Annual ACM Symposium on Applied Computing, SAC 2018
T2 - 33rd Annual ACM Symposium on Applied Computing, SAC 2018
Y2 - 9 April 2018 through 13 April 2018
ER -