基于AHP-Apriori算法的事故成因研究

发布时间：2018-07-17 21:07

【摘要】：交通安全问题已成为世界范围内共同关注和探究的一大难题，交通事故的发生是突发、偶然的，受各种因素共同支配，是各影响因素相互作用的综合结果。如果能够发现蕴含在事故数据中的规律，如道路条件、车辆状况、时间、天气与事故本身属性中的事故类型、事故形态等的相互作用，提前做好安全预防措施，将事故发生的风险控制在可控范围内，就能有效防止事故的发生，减少其带来的危害。目前各地道路交通管理部门对事故数据进行了详尽的记录，建成了交通事故信息系统，但是仍然应用以统计学手段为主的传统概率回归等分析方法忽视了更具决策意义的交通事故数据整体特征描述、数据内部之间的相关关系、交通事故发生规律及发展趋势的预测。使得采集到的交通数据并没有得到充分的挖掘和应用，冗杂的原始数据占用大量的存储空间，造成资源浪费。论文基于公安部道路交通事故数据库，采用数据挖掘技术，根据交通事故成因多维度、层次化的特点，从驾驶员、车辆、道路、环境四大要素出发，在客观分析基础上将其量化，对交通事故成因进行研究；构建基于AHP-Apriori算法的事故模型，首先利用层次分析法（AHP）对事故成因进行重要度排序，找出事故影响主因素；在此基础上，采用关联规则算法对筛选出的主要影响因素进行事故关联分析。分析表明，该方法能够确定不同条件下的各事故因素组合危险性及事故发生的时空特性，特别是在大交通事故数据样本的情况下，能提高运算效率和分析的准确性。最后基于该算法对道路安全水平进行安全评价和预警，并提出安全管理建议和道路安全措施，用以指导交通管理部门决策，杜绝事故隐患、减少事故发生，，最终提高道路的安全水平。
[Abstract]:The problem of traffic safety has become a major problem which is concerned and explored in the world. The occurrence of traffic accidents is unexpected, accidental, and controlled by various factors, which is the comprehensive result of the interaction of various factors. If we can find the interaction between the rules contained in the accident data, such as road condition, vehicle condition, time, weather and accident type in the attribute of the accident itself, the accident form, and so on, we can do a good job of safety prevention measures in advance. If the risk of accident is controlled, it can effectively prevent the accident and reduce its harm. At present, local road traffic management departments have carried out detailed records of accident data, and built a traffic accident information system. However, the traditional probability regression analysis method, which is based on statistics, still ignores the overall feature description of traffic accident data, which is more decision-making significance, and the correlation between the data. The law of traffic accident occurrence and the forecast of development trend. The collected traffic data are not fully mined and applied, and the jumbled raw data occupy a large amount of storage space, resulting in a waste of resources. Based on the road traffic accident database of the Ministry of Public Security, this paper adopts data mining technology, according to the multi-dimensional and hierarchical characteristics of the cause of the traffic accident, starting from the driver, the vehicle, the road and the environment. On the basis of objective analysis, the causes of traffic accidents are quantified, and the accident model based on AHP-Apriori algorithm is constructed. Firstly, AHP is used to rank the importance of the causes of accidents to find out the main factors affecting the accidents. Association rules algorithm is used to analyze the main influencing factors. The analysis shows that this method can determine the combination risk of various accident factors under different conditions and the space-time characteristics of accidents, especially in the case of large traffic accident data samples, which can improve the calculation efficiency and accuracy of the analysis. Finally, the road safety level is evaluated and early warning based on the algorithm, and safety management suggestions and road safety measures are put forward to guide the traffic management department to make decisions, to put an end to the hidden trouble of the accident, and to reduce the occurrence of the accident. Eventually improve the safety level of the road.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.31

【参考文献】