Core Passenger Identification Of Rural Customized Bus In Haining City Based On Similarity And Repeatability

Nowadays,conventional public transport is facing the difficulty of less attractive and passenger capacity.Many cities carry out customized public transport services to promote the development of urban public transport system.However,these cities have not been accurately grasped the characteristics of passenger travel demand.Haining City plans to build customized public transport services,which is facing the same problem.With the development of big data technology,the correlation between passenger travel data can be found through the analysis of smart card data.Precisely identify the core group of passengers who have a high demand for customized buses.The core passenger in this paper refers to the passenger group with the most specific long-term travel mode.Based on the bus card data in September 2021,this paper regards the rural area as rural passengers swiping their cards for the first time every day,and selects the passengers who have more than one week of frequent bus trips each month.The rural IC card(inter-city card,senior card and free senior card)and mobile payment are 1189,1172,756 and 1171,respectively.In this way,the core passengers of customized bus in rural areas can be identified.Firstly,bus passenger travel stations are clustered and the whole card swiping time is taken by hour,so as to reconstruct time and space dimension travel.The T-O-D space-time travel mode of bus passengers was constructed by single trip.The spatial and temporal travel patterns were used to measure the travel similarity between passengers by cosine similarity function.Then,the similarity value was used to classify the bus passengers.The number of classified clusters was determined by contour coefficient,and the clusters of different similar levels were divided.Then,according to the number of travel modes of passengers,the stickiness coefficient method is used to measure the repeatability intensity of passengers’ travel modes in the cluster.The final excavation viscosity coefficient value above 0.5 for customized bus core passengers.Finally,the travel characteristics of the core passenger groups of the two kinds of payment are compared and analyzed,so as to put forward the group application management suggestions.The results showed that:(1)The clustering effect of passenger group is good.The contour coefficient ofeach payment is not less than 0.8.(2)87%,63%,68% and 64% of the high similarity clusters of mobile payment,intercity card,senior card and free senior card passengers were core passengers,with high travel repeatability intensity.Finally,63,217,130 and 88 core passengers were identified and excavated respectively.(3),Travel difference: in the spatial dimension,the intercity core commuting cluster with card is mainly intra-town travel,while the mobile payment cluster with senior card is mainly urban-rural travel.In the temporal dimension,the morning peak is 6:00 for three IC cards and 7:00 for mobile payment.The results show that:(1)According to the passenger mining method combining similarity and repeatability,this paper has a good effect on identifying the core passengers of customized bus in Haining City.(2)The travel characteristics of core payment groups are significantly different,so customized bus services need to be more targeted;(3)By comparing the travel characteristics of various payment methods,among them the passenger group mining effect of intercity card is better than that of mobile payment,senior card and free senior card,and it is the main group of rural customized bus in Haining City.