Research On The Prediction Method Of Public Bike Demand Based On The Cluster Of Stations

Public bicycle,As a green means of transportation,is very popular among people and it is an important way for many people to travel daily.In recent years,with the increasing scale and frequency of bike sharing system,the unbalanced use of shared bicycle has a great impact on the users’ experience,which makes the use rate of public bicycle not maximized.However,the existing research methods have the problems of incomplete data consideration and lack of feature processing before prediction,which makes the prediction accuracy of the method not very high and has certain limitations in applicability.In view of the above problems,this paper proposes a method to forecast the usage demand of public bicycle based on the cluster of stations.Firstly,the temporal and spatial distribution of public bicycles is analyzed by integrating the data of public bicycles and meteorological data,and the relevant characteristics affecting the demand of public bicycles are extracted.Secondly,using the extracted features,the paper proposes a method combining k-medoids clustering,association rules and total usage demand constraint adjustment(optimizing clustering to achieve the maximum balance)of the cluster to cluster the stations with similar usage patterns of public bicycles,so as to improve the prediction accuracy of public bicycles.Finally,on thebasis of the station cluster,we get the characteristics of cross cluster usage,and then combined with meteorological and time characteristics to build Elman neural network prediction model based on the cluster of stations.This paper uses dataset from the New York City public bicycle system to evaluate our method.The feature selection proves that the feature extracted in this paper is very helpful to improve the performance of the prediction method.At the same time,by comparing with RF,BP and HP-MSI baseline methods,it also proves that our method has the advantage of exceeding the baseline method in accuracy.Our method can not only give full play to the utilization rate of public bicycles and solve the imbalance of distribution of public bicycles,but also provide theoretical and practical basis for the planning,layout and scientific scheduling of shared bicycle system.