面向道路交通违法行为的事理图谱构建技术

我国道路交通安全法律法规体系不断完善，要求交警部门针对具体的交通违法行为给予不同程度的处罚，为了响应国家以智慧化推动治理现代化的号召，可利用事理图谱技术构建道路交通领域知识库，揭示道路交通违法行为事件之间的逻辑关系，能快速且有效辅助处理交通违章事件。以开源数据为基础，面向道路交通违法行为构建事件语料库，通过事件抽取、关系抽取等步骤，构建交通违法事理图谱，在此基础上，实现了交通违法问答系统。研究结果表明：利用所提模型抽取道路交通违法行为事件的分类评价指标F1值达到0.832，识别率很高。此外，基于事理图谱的交通违法问答系统展示了事理图谱技术在道路交通领域具备良好的应用前景。

Development of a Knowledge Base for Reasoning Penalty for Traffic Violations Based on Event Evolutionary Graph

With continuous improvement of laws and regulations related to road safety in China, traffic police departments are required to issue different penalties for specific traffic violations. In response to the call and to improve the capacity of"intelligent governance", this article proposes to develop a knowledge base for traffic violations and accidents with event evolutionary graph, which can be used to reason appropriate penalty for traffic violations& accidents quickly and efficiently. This paper uses open-source data to develop the knowledge base required for processing traffic violations/accidents and creates an event evolutionary graph through extracting traffic events and their relationship. Moreover, a knowledge base system for traffic violations/accidents is developed. The experimental results show that the proposed system offers a F1 score of 0.832 when classifying traffic violations and accidents, which indicates that the event evolutionary graph is a good tool for reasoning the penalty of traffic violations and accidents.