A Study Of Trip Generation Intensity Forecasting Method Based On Land Use Information

Trip generation intensity is often used to estimate the trips generated in the first step of the traditional four-step travel demand forecasting method,which can be a good representation of trip characteristics of residents.The trip generation rate is an important measure of trip intensity of residents.It is used to characterize the unit trip intensity of a certain type of land use,and is composed of two parts: trip production rate and trip attraction rate.However,since the accurate land use information is often hard to obtain,it is difficult to know exactly where travelers originate and where they go.The forecasting results,therefore,are often questioned about their accuracy in practice.With the advent of the era of big data,the Internet has become the main source for people to obtain information.Web information extraction technology makes it possible to obtain refined and routine updated land use information in the Internet.Aiming to address the problem of trip generation rate study,a new method is proposed to obtain the traffic analysis zone(TAZ)point of interest(POI,representing any meaningful location such as a store,school,hospital,residential district,etc.)information from web information extraction to represent detailed land use information and to establish multiple regression forecasting model and GWR model of trip generation.The model is then calibrated and validated in a case study using the OD trips estimated through mobile phone data in Chaoyang District,Beijing.A comparison of the trip generation results shows that the mean absolute error of the trip production estimation from the mobile phone data and the results from the proposed method is about 14%,and the mean absolute error of the trip attraction estimation from the mobile phone data and the results from the proposed method is about 13%.The coefficients of the model provide alternative estimates of trip generation rates.The proposed method is an exploration of data mining and web information extraction techniques in traditional transportation planning.It will not only help to reduce the workload of data extraction,but also help researchers and planners to obtain more detailed land use information,and therefore,improve the travel demand forecasting accuracy.