In a crash report, crash narrative is used to describe the sequence of events for all units involved in the crash, and record additional information. As every crash scene contains unique aspects or circumstances, the narrative description of observed events provides irreplaceable and crucial information that cannot be captured in the structured data fields. Text mining and machine learning techniques have been proven to be efficient and effective in automatically extracting crucial information from crash narrative to facilitate crash analysis and crash classification, particular for the ones that have been misclassified or overlooked. Our goal is to develop an online Crash Information Extraction, Analysis and Classification Tool (CIEACT). The engine of the tool is the models developed from NoisyOr classifier and the neural network model GRU. The interface of the tool will be an interactive crash map that can display the results and support safety analysis in a spatial context. The functions and analysis offered by CIEACT can provide safety practitioners and professionals with maximum and quick access to information stored in the texts of crash narrative and substantially reduce crash report review time.
Use the tool: CIEACT Tool
Project Documents
Project Details
Project ID
MIL117752
Status
Complete
Start Date
April 1, 2021
End Date
September 30, 2021
Focus Areas
Data Analytics, Modeling and Simulation
Safety
Sponsors
US Dept Of Transportation
Wi Dept Of Transportation
Research Centers
Institute for Physical Infrastructure and Transportation (IPIT)
Principal Investigator
Qin, Xiao
Director, Institute for Physical Infrastructure and Transportation (IPIT)
Professor, Civil and Environmental Engineering University of Wisconsin-Milwaukee
Co-Principal Investigator
Kate, Rohit J
Associate Professor Department of Computer Science University of Wisconsin-Milwaukee