A major focus of traffic network analysis is to identify important road sections and intersections in the region which contributes to effective traffic intervention and has important practical significance in traffic breakdown research. By modeling the actual road network and analyzing it, important nodes can be identified according to the evaluation rules and critical road sections and intersections can be determined. On the basis of traditional network research methods, this paper refers to a biomedical research called “Network-based global inference of human disease genes” which developed a tool named CIPHER to predict and prioritize disease genes. This paper uses CIPHER to rank nodes in the traffic network instead of ranking genes. First, a data set is established through the simulation of traffic breakdown on the actual traffic network in Langfang city. The characteristics of traffic condition and traffic breakdown are described by vectors refers to WordEmbedding. The traffic condition of road network is regarded as phenotypic network. Second, matrices represent the topological distance between road network nodes which simulates the gene network. Third, CIPHER is used to give the consistency score of certain nodes and rank them. The score is a linear correlation coefficient to measure the consistency between the position of certain node in the traffic network and the variations of phenotypic similarity for the traffic state in the whole traffic breakdown data set network. Then the evaluation results of nodes using CIPHER is compared with those using traditional methods. The selected important nodes and analysis results show the feasibility, advantages and disadvantages of CIPHER in traffic breakdown analysis and warning. CIPHER links the actual state of the road network to the data set of traffic breakdown that have occurred and can be used to judge whether the node is crucial in traffic breakdown and recovery.

CICTP