Traveler-Assisted Information Extraction from

초록

Traffic problems make the daily travel inconvenience in urban areas. People need an intelligent system that notifies them before involving in such a problem. Therefore, this paper proposes a Convolution Neural Network (CNN)-based information extraction and sentiment analysis system. The proposed approach retrieves a real-time information related to traffic problem (road flooding, traffic jams, etc.) from social network, preprocesses this information in order to filter out irrelevant data, applies CNN to predict sentiment of transportation tweets, and then uses ontology to provide transportation features along with polarity to travelers.