Forecasting Traffic land Demand in Guangdong-Hong Kong-Macao Greater Bay Area Based on Gray-BP Neural Network Model

Abstract

In addition to promoting rapid regional socio-economic development, traffic land will occupy a large amount of land resources, leading to conflicts between different land types. Therefore, it is necessary to forecast the traffic land demand with scientific methods. This paper takes the Guangdong-Hong Kong-Macao Greater Bay Area as a case study to forecast its traffic land demand for the next 8 years. The data were obtained from the regional yearbook from 2008 to 2020 and the shared application service platform of land survey results of the Ministry of Natural Resources. A reasonable index system of impact factors was established according to its development characteristics, and a gray correlation model was used to rank the importance of impact factors, and finally a coupled gray-BP neural network model was constructed to forecast the traffic land demand. The forecasting error of traffic land demand of 11 cities in Greater Bay Area from 2021-2028 obtained under this method is about 0.1%.