Newly-coined words and emoticons dictionary construction for social data sentiment analysis

Abstract

In the composition of social networks and social networks services, most of articles are written with writer's personal ideas and emotions within 200 words. Social networks services produce social big data and provide vast volumes of data. Social big data has various types of text-centric data, photos, music and videos, and informal data. It is difficult to deal with social big data, since it has rapid changes and volume growth. But, analysis of social big data may accept the online market trust agreement and establish national policies to support the marketing and strategy formulation of the company for marketing and strategy establishment. This paper proposes a new social big data analysis system that uses new-coined words and emoticons for social big data analysis. In this paper, we also construct a new dictionary of emoticons to improve the accuracy of big data analysis related to the public opinion, and to improve the accuracy of emotional analysis using the existing emotional dictionaries and the newly constructed new emoticons and emoticon dictionaries at the analysis stage. Especially, the accuracy of social big data analysis could be improved by quantifying and analyzing emotional level about emoticons and new-coined words. This research built a dictionary of emoticons, and used the Naver open dictionary for newly-coined words. But, newly-coined words that were not included in the Naver open dictionary are added to proposed newly-coined words database.