A Novel Development Trend Prediction Approach of Digital Financial Inclusion Based on Exploratory Spatial Data Analysis and Graph Convolutional Neural Network

We proposed a novel development trend prediction approach of digital financial inclusion (DFI) based on exploratory spatial data analysis (ESDA) and graph convolutional neural network (GCN), named ESDAGCN. Firstly, to address the limitation of standard GCN in lacking financial background knowledge, the construction of an adjacency matrix is enhanced by incorporating global Moran's I analysis. Secondly, the Moran matrix is obtained in the light of spatial dependency, and it is utilized to construct adjacency matrix. Finally, ESDAGCN is wielded to explore the spatial dependency of the development of DFI. The proposed model provides a new idea for DFI research