Investigation on Spatial Spillover Effect of Transportation Infrastructure on Regional Economy in Yangtze River Economic Belt of China

Abstract

Exploring the mechanism of the spatial distributed impacts of transportation systems on the local economy helps realize the coordinated development of the local economy and environmental protection. Therefore, correlation tests are conducted using linear regression models, decision tree regression models, machine learning random forest regression models, and spatial econometric models in this study. An extensive analysis was carried out to determine the relationship between the transport infrastructure of the Yangtze River Economic Belt and the impact on regional economic mechanisms and transmission paths from the spatial spillover effect. The three models are compared and analyzed to select the best model. Then the Moran index is calculated to study whether the data have spatial correlation and spatial difference. After confirming the spatial impact of the data, the Lagrange Multiplier test, the Likelihood-Ratio test and other relevant tests were used to select the spatial econometric model that best fits the data set. Finally, the study results show that the random forest model has a higher R² of 95.14%. To analyze spatial economics, it is recommended to use the fixed effect and double fixed effect of the Spatial Durbin Model (SDM). Road and Density of the transportation network explain GDP more significantly.