Evaluating model of traffic accident rate on urban data

Abstract

Abstract-Public safety, especially the daily traffic accident is concerned by the public. Previous studies have already discussed accident reasons associated with accidents statistically. There is a method called Innovators Marketplace on Data Jackets created by Professor Ohsawa. This method is used to externalize the value of data via stakeholders' requirement communication. This paper applied the solution from an IMDJ workshop to research this topic creatively. This novel solution suggested to do analysis on the combination of urban data and traffic accident rate to find the impact factors to the traffic accident rate in the urban system. This paper used factor analysis, structure equation modeling and data mining to construct a theoretical frame for traffic accident rate analysis for urban data. Different accident indexes, such as total number of accident, fatality rate, injury rate, and casualty rate are combined to construct a traffic accident risk evaluation model. This paper chosen the urban data as the solution from IMDJ workshop, such as population structure information, vehicle information, road characters, public traffic system information, and the other kinds of data to explore factor meaning, and to identify relationships between different factors. It segmented these urban data based on their categories, and determined accident risk for each section. By doing analysis on not only the original data but also the changing rate of these data each year, the result analytical results showed that traffic accident rate on urban data could be described by the combination of population structure, road characters, public traffic system and public facilities. These four sections affects traffic accident rate significantly during the development of urban; however, the vehicle factor does not have influence on traffic accident rate. And it proves the solution from IMDJ workshop is not only novel but also practical strongly. Making some solution from IMDJ into reality, we will find another new way to affect the world.

A Machine Learning Approach for Classifying Road Accident Hotspots

“…Wang and Ohsawa [26] proposed an evaluation model for traffic accident risk, where they defined the relationship between urban data and traffic accident data. The authors used data mining to build a framework for analyzing the traffic accident rate using urban data from Beijing, China.…”

Los accidentes de tránsito desde la perspectiva de la minería de datos. Una revisión de la literatura.

“…Volumen 8, Número 2, Mayo -Agosto de 2020, ISSN 2346-030X PP 133-141. Unos de los temas más tratados por los investigadores son las causas externas [16] [20][21][22][23][24], estas son analizadas en países como India, China, Estados Unidos y otros. Se destacan dos artículos de la India que analizan accidentes donde estuvieron involucrados los vehículos más comunes en ese país conocidos como "auto-rickshaws" o "tuc-tuc" [25] y las motocicletas [26].…”