Extraction of Trend Keywords from Thai Twitters using N-Gram Word Combination

Abstract

Extracting keywords from text on social media facilitates people to update news and trends. It reduces time spent for identifying main content from huge amount of data, and it can be used to identify situations or events that most of people mention in each period of time. This paper proposes a method for extracting keywords from Thai text on social media. A N-gram-based word-combination technique is presented to segment words that are not in dictionaries and increase the precision of word segmentation. Posts on Twitter concerning universities in Thailand are used as a case study for extracting keywords and analyzing trends. The experimental results show that the proposed method yield the highest precision of 70%.