Traffic information extraction and classification from Thai Twitter
Supon Klaithin, Choochart Haruechaiyasak · 2016
Twitter is a one of the most popular microblogging service. It is broadly used in communication. Besides, it provides up-to-date real-time traffic information source. Various information contained in tweets are such as accident, road name, and place name. In this paper, we extract and classify tweets by tagging traffic information. They were extracted into 12 tags and classified into 6 categories. Our study indicate that 3 significant components were information of road, information of location, and traffic status. Furthermore, the classification accuracy achieved with the testing data was 76.4%. The average accuracy of information extraction is about 88.42%. The two high accuracy of information extraction were location (LOC) and time (TIM), which found that nearly 99.42% and 97.09% of all categories respectively.