Exploring traffic accident locations from natural language based on spatial information retrieval

Exploring traffic accident locations is essential for making prevention strategy in order to improve traffic safety proactively. Quite a number of methods have been developed to acquire accident spatial data, by various detectors in roads. However, plenty of locations information is contained in textual alarm information transformed from telephone calls of 122 Alarm Reception. This paper, aims to obtain spatial data, longitude and latitude, of traffic accident locations based on textual alarm information. Natural language processing, database system and Geographic Information System (GIS) are used to get traffic accident spatial data. The proposed method involves three processes, location name identification, spatial information retrieval and information matching respectively. Validation results show that the method is of reasonably high accuracy for exploring traffic accident locations.