Estimating the polarity index of a word
- Authors
- Park, H.; Lee, S.
- Issue Date
- Aug-2017
- Publisher
- Medwell Journals
- Keywords
- Experimental results; Korean dctiomq information; Polarity index estimation; Sentiment analysis; Sentiment dictionaq; Sentiment word
- Citation
- Journal of Engineering and Applied Sciences, v.12, no.16, pp.4022 - 4027
- Journal Title
- Journal of Engineering and Applied Sciences
- Volume
- 12
- Number
- 16
- Start Page
- 4022
- End Page
- 4027
- URI
- http://scholarworks.bwise.kr/ssu/handle/2018.sw.ssu/7359
- DOI
- 10.3923/jeasci.2017.4022.4027
- ISSN
- 1816-949X
- Abstract
- In sentiment analysis, sentiment dictionary is important because it affects the results of sentiment analysis. To build a sentiment dictionary, it is needed to classify the sentiment category of a word and to quantify its sentiment polarity strength. In English-based studies, it is easy to build an English sentiment dictionary to quantify the sentiment polarity strength of a word. However, such English sentiment dictionary is unsuitable for sentiment analysis of Korean documents. In Korean based studies, it is difficult to perform sentiment analysis for Korean documents precisely because most of Korean sentiment dictionaries with the sentiment polarity strength are built by surveys and the number of words is limited. In this study, we propose a method to automatically estimate the polarity index of a word using the synonyms and definitions information of words in a Korean dictionary. The experimental results show that the proposed method outperforms methods of using only one of the synonyms or definitions information.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Information Technology > School of Software > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/ssu/handle/2018.sw.ssu/7359)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.