Abstract—The purpose of this study is to construct the spatial characteristics of pedestrians' accident points as learning materials and to develop a model for deep learning them and to predict the risk of similar areas. In order to construct learning materials for pedestrian accident points, the characteristics of geometric factors related to pedestrian accidents at the accident points were quantitatively analyzed through spatial analysis. As a result of the study, in the analysis of the deep learning pattern, the prediction results could be improved by changing the middle layer and adjusting the parameters. It is expected that this result will contribute to lowering the accident rate of pedestrians by preparing measures for high accident rate points by reviewing vulnerable traffic accident sites in the future.

Index Terms—Spatial analysis, deep learning, transportation data, pedestrian accident.

The authors are with Department of GIS Engineering, Namseoul University, Republic of Korea (email: ilyoung.hong@nsu.ac.kr).

Cite: Ilyoung Hong, Hanseung Choi, Songpyo Hong, "Risk Assessment of Pedestrian Accident Area Using Spatial Analysis and Deep Learning," International Journal of e-Education, e-Business, e-Management and e-Learning vol. 8, no. 4, pp. 248-256, 2018.