User Data Mining And Application Research Of Public Bicycle System

In recent years,with the dramatic increase of population,it brings a series of problems to the city.In the face of these problems,the public bicycle system is becoming more and more popular in the modern transportation system.Public bicycle system provides convenience at the same time also brings a series of challenges to the system operators:some stations of location or are set unreasonably;users sometimes are unable to return or rent bicycles;damage rate of bicycle is very high.In the running process of the system,a large number of user trip data have been accumulated,on the basis of the analysis of these data,a series of applications are proposed are of great significance to the development of the system.The historical data of public bicycle is a kind of spatial and temporal data,and spatial-temporal data mining is a kind of implicit,valuable and new knowledge from massive,multi-dimensional and non-linear data mining.In this paper,the Washington public bicycle system as the research object,we propose two applications of bicycle of balanced use and lease points function identification based on the user trip data mining,which can effectively reduce the loss rate of the bicycle and provide theoretical and applicable support for the system layout optimization and scheduling,so as to improve the user satisfaction and system efficiency.This paper main talk about following questions:First of all,we analyze the historical user trip data of the system and the time and space characteristics of public bicycle system are discussed,which can prepare for the follow-up study.Secondly,we put forward the application of bicycle balance.This paper is inspired by the Cache replacement strategy in the computer system structure,propose four kinds of bicycle equilibrium models,and the validity of the four models is analyzed under different replacement criteria based on the bicycle lifetime model.In the end,we present the application of the public bicycle system's lease points function identification.We employ LDA(Latent Dirichlet Allocation)model and K-means algorithm to realize the function of identifying rental points;use the usage feature of people,POI(Point of Interest)data and stations' information for result analysis and description.