Convolutional neural networks (CNN) have achieved excellent performance for the hyperspectral image (HSI) classification problem due to better extracting spectral and spatial information. However, CNN can only perform convolution calculations on Euclidean data sets. To solve this problem, recently, graph convolutional neural network (GCN) is proposed, which can be applied to the semi-supervised HSI classification problem. However, the GCN is a direct push learning method, which requires all nodes to participate in training process to get the node embedding. This may bring great computational for the hyperspectral data with large number of pixels. Therefore, in this paper, we propose a inductive learning method to solve the problem. It constructs the graph by sampling and aggregating (graphSAGE) feature from a node's local neighbourhood. This could greatly reduce the space complexity. Moreover, to enhance the classification performance, we also construct graph using spectral and spatial information (spectral-spatial GraphSAGE). Experiments on several hyperspectral image datasets show that the proposed method can achieve better classification performance compared with state-of-the-art HSI classification methods.

As a new driving force for economic growth, the digital economy has had a profound impact on the labor market. While the existing research has explored the role of the digital economy in job substitution, creation, and polarization effects, the research on the impact on the social insurance fund income is relatively scarce. In view of this, based on the provincial panel data from 2011 to 2020, this paper analyzes the effect and mechanism of the digital economy on the pension income by using the moderated chain intermediary model and random forest regression. The results show that: (1) the employment scale, labor income, industrial structure, and government intervention are the important factors affecting the income of urban pension insurance; (2) the development of the digital economy has a negative impact on the income of the basic pension insurance fund for urban employees, and the chain intermediary effect that indirectly affects the employment scale and labor income through promoting the upgrading of the industrial structure has a negative impact on the income of the pension insurance fund. The employment scale and employment income of the industries with high and low substitution rates have a significant impact; (3) government intervention can regulate the negative impact of the digital economy development on the pension fund income. Furthermore, taking the transformation and reform of social security collection and payment institutions in July 2018 as an opportunity, the analysis using the event study method found that the average level of the pension income in the regions where the tax department was fully responsible increased significantly compared with the regions where the social security department collected it. Therefore, in order to maintain the sustainability of the pension fund income and effectively prevent the problem of old-age poverty caused by the “silver wave” and the lack of protection of workers’ rights and interests, institutional innovation should be promoted, the current tax policy should be adjusted, and the inclusiveness and flexibility of the pension security system should be improved. Digital technology should be used to improve the government’s intervention capacity and management level, and promote the positive interaction between the digital economy and the pension insurance system.