The state subdivision of public traffic vehicles based on K-means algorithm

With the rapid development of public transportation, the operating intensity of public transport vehicles has been continuously increasing and various kinds of sudden problems have appeared in high-intensity and overloaded buses. This article aims at the problem of bus group vehicle fault status, analyses the vehicle maintenance and repair data and combines the weather data to statistically analyse the fault information. From the viewpoint of vehicle value, the value of nearness, frequency and time are selected as indicators of vehicle state breakdown. The cluster analysis of bus vehicles is performed using the K-means clustering method, which is divided into high fault cars and low fault cars. Three different groups of general fault car labels and the use of the profile coefficient to verify the results of cluster analysis, it is proved that the classification results have better incremental self-learning ability and level of cognitive ability, help to find and solve problems in advance. Different maintenance strategies are formulated to reduce the number of bus breakdowns and reduce rescue costs.