关键词: 特殊事件, 中美贸易战, 国际关系, GDELT, 复杂网络, 网络演化, 动态社区划分, 空间分析

Abstract:

Since the beginning of the 21^st century, the complex and fast-changing international relations have brought profound effects to the world economy, politics, security, and diplomacy. Keeping abreast of the changes in international relations is of great significance to China's foreign policy making and overall development planning. With the advent of the era of big data, the application of big data combined with quantitative analysis tools of international relations to timely and effectively mine the change patterns of international relations has become an important issue. Big data of news events with strong timeliness and high information content can timely reflect the information that international events affect global international relations. As an information mining method oriented to big data, network mining is a good choice for data-driven international relations research because of its figurative relational expression and rich structural analysis methods. Taking short-term international events as the background, the network mining of big data of news events and the sequential evolution analysis of international relations network can provide solutions to the changes of international relations in the context of changes of international relations caused by short-term international events. This paper takes the Trade War between China and the United States as an example to study the temporal evolution pattern of the international relations network in special events. Based on the GDELT news event data, the international relations network is constructed, and methods based on complex network theory are used for information mining and analysis of international relations. Firstly, the data are used to construct the international relationship network, then the temporal evolution patterns are detected by dynamic community partition method, and finally the spatial characteristics are analyzed by combining the spatial analysis methods such as analysis of point distribution patterns, nuclear density analysis, and spatial autocorrelation analysis. The results show that: (1) In the process of the occurrence of special events, there is a strong correlation between the evolution of the network community and the type of sub-events; and (2) Nodes in the same community generally show obvious clustering characteristics in spatial distribution. Nodes in a specific region join different communities with high frequency. The spatial distribution of the high values of node network attribute changes with the occurrence of events. The local eigenvalues of the network change dramatically with the occurrence of sub-events. The research in this paper provides a new perspective for the empirical analysis of the dynamic changes of international relations in short-term international events, provides a new idea for the spatial shift of international relations research, complements the data-driven international relations research at the methodological level, and also provides a reference for the network mining of big data.

Key words: special events, China-US trade war, international relations, GDELT, complex networks, network evolution, dynamic community division, spatial analysis