JUCS - Journal of Universal Computer Science 18(8): 973-992, doi: 10.3217/jucs-018-08-0973
Discovering Consumer Insight from Twitter via Sentiment Analysis
Wilas Chamlertwat‡, Pattarasinee Bhattarakosol‡, Tippakorn Rungkasiri‡, Choochart Haruechaiyasak§‡ Chulalongkorn University, Bangkok, Thailand§ National Electronics and Computer Technology Center, Pathumthani, Thailand
Corresponding author: Wilas Chamlertwat ( wilas@chamlertwat.in.th ) © Wilas Chamlertwat, Pattarasinee Bhattarakosol, Tippakorn Rungkasiri, Choochart Haruechaiyasak. This article is freely available under the J.UCS Open Content License. Citation:
Chamlertwat W, Bhattarakosol P, Rungkasiri T, Haruechaiyasak C (2012) Discovering Consumer Insight from Twitter via Sentiment Analysis. JUCS - Journal of Universal Computer Science 18(8): 973-992. https://doi.org/10.3217/jucs-018-08-0973 |  |
Abstract
Traditional approaches for studying consumer behavior, such as marketing survey and focus group, require a large amount of time and resources. Moreover, some products, such as smartphones, have a short product life cycle. As an alternative solution, we propose a system, the Micro-blog Sentiment Analysis System (MSAS), based on sentiment analysis to automatically analyze customer opinions from the Twitter micro-blog service. The MSAS consists of five main functions to (1) collect Twitter posts, (2) filter for opinionated posts, (3) detect polarity in each post, (4) categorize product features and (5) summarize and visualize the overall results. We used the product domain of smartphone as our case study. The experiments on 100,000 collected posts related to smartphones showed that the system could help indicating the customers' sentiments towards the product features, such as Application, Screen, and Camera. Further evaluation by experts in smartphone industry confirmed that the system yielded some valid results.
Keywords
micro-blog, sentiment analysis, natural language processing, information extraction, information visualization