Real-time road transportation safety risk evaluation model based on data-mining

Abstract

Aiming at the shortcoming of the massive amount of manual work required and the result relay on the cognitive level of experts the process cannot be implemented dynamically. A real-time road transportation safety risk evaluation model based on data-mining is proposed in this paper. Firstly, the pre-processing of unstructured text contains process of adding custom item dictionary, deletion of stop words, word segmentation of text at the beginning process, then dynamic risk factors identification using TF-IDF on processed text. Secondly, accident chains extraction by cue words and causal sentence structure construction. Thirdly, the relevance mining among risk factors or accident states through Apriori-algorithm. Finally, real-time risk assessment is realised by classification of the product of obtained probability and severity degree result using K-means. Experiments are conducted on text data set, and the result shows that the accuracy of proposed model is 88%, which is an effective safety risk evaluation model.