Traffic Condition Information Extraction From Twitter Data

Abstract

Twitter is a social media that allows users to send various information including information about traffic conditions. This information can be utilized to map the traffic condition in the city at a certain given time. Research regarding traffic information extraction in Bahasa Indonesia has been done previously by using a rule-based method. This method has a weakness that is required more effort to make the rules and the Out of Vocabulary (OOV) and Out of Rules (OOR) problem. In this research, we propose an information extraction methods to map a traffic condition from a tweet. This method implemented by tweet classification, Named Entity Recognition (NER), and relation extraction. In those three methods, we utilize a supervised machine learning algorithm, Support Vector Machine (SVM). The results of this study indicate that the information extraction system which consists of tweet classification, NER, and relation extraction were capable to extract traffic condition information from Bahasa Indonesia tweet.