Detecting Opinion Polarities using Kernel Methods

Rasoul Kaljahi, Jennifer Foster


Abstract
We investigate the application of kernel methods to representing both structural and lexical knowledge for predicting polarity of opinions in consumer product review. We introduce any-gram kernels which model lexical information in a significantly faster way than the traditional n-gram features, while capturing all possible orders of n-grams n in a sequence without the need to explicitly present a pre-specified set of such orders. We also present an effective format to represent constituency and dependency structure together with aspect terms and sentiment polarity scores. Furthermore, we modify the traditional tree kernel function to compute the similarity based on word embedding vectors instead of exact string match and present experiments using the new models.
Anthology ID:
W16-4307
Volume:
Proceedings of the Workshop on Computational Modeling of People’s Opinions, Personality, and Emotions in Social Media (PEOPLES)
Month:
December
Year:
2016
Address:
Osaka, Japan
Editors:
Malvina Nissim, Viviana Patti, Barbara Plank
Venue:
PEOPLES
SIG:
Publisher:
The COLING 2016 Organizing Committee
Note:
Pages:
60–69
Language:
URL:
https://aclanthology.org/W16-4307
DOI:
Bibkey:
Cite (ACL):
Rasoul Kaljahi and Jennifer Foster. 2016. Detecting Opinion Polarities using Kernel Methods. In Proceedings of the Workshop on Computational Modeling of People’s Opinions, Personality, and Emotions in Social Media (PEOPLES), pages 60–69, Osaka, Japan. The COLING 2016 Organizing Committee.
Cite (Informal):
Detecting Opinion Polarities using Kernel Methods (Kaljahi & Foster, PEOPLES 2016)
Copy Citation:
PDF:
https://aclanthology.org/W16-4307.pdf
Data
MPQA Opinion Corpus