Analysis of the factors influencing highway crash risk in different regional types based on improved Apriori algorithm

There are many factors affecting the risk of highway crash, and they are related to each other. To figure out the factors relating to crash risk in different regional types and their inner relation links, so that it contributes to the traffic safety in highway, this study takes three sections of highway (areas of downtown, suburb and mountain, in Washington State, USA) as the research object, and data was collected in accordance with the five elements dynamic system of “people - car - road - environment - management”. To improve the operation efficiency and highlight the main association rules, Analytic Hierarchy Process (AHP) was applied. Based on AHP improved Apriori association rule mining algorithm, crash risk influencing factors and their complex association rules were identified. The result shows that in the downtown area, the values of support degree, confidence degree and lift degree in the ranking top 10 support degree association rules are 0.87 to 0.93, 0.93 to 1.00, 1.00 to 1.02 separately, and in the ranking top 10 support degree association rules are 0.32 to 0.33, 1.00, 2.67 separately. Under a lower support degree and higher confidence degree situation, the value of top three rules with highest lift degree are all 3.75. Case study shows the method adopted is operative. Meanwhile, highway in different regional types has different crash occurrence mechanisms, and the same factor or association rule has different values in different regional types.