Risk Perception of the " Belt and Road" Countries Based on Global Media Data GDELT

Abstract

Rapid and dynamic perception of risks in countries along the “Belt and Road” is of great practical significance for the construction of the “Belt and Road”. The current measurements of country risk are divided into two categories: multi-factor analysis and systematic risk modeling based on capital asset pricing theory. There are problems such as time lag in data updates and inadequate completeness. Risk perception based on big data has the characteristics of wide sources, high timeliness, multiple dimensions, and full coverage, and it can capture potential risk variations earlier and faster. In this study, based on the global media big data GDELT, it is found that the risks of the countries along the “Belt and Road” are mainly focused on politics, military risks, energy trade, terrorism, power struggles, etc. West Asia and North Africa region are at the core of the risk reports in the network of countries mentioned, with the Central and Eastern Europe and Central Asia region playing the role of "bridge" nodes. In the “country-topic risk” heterogeneous information network, when the risk topic similarity was set to 0.7, the country risk clustering effect is the best. Syria had always been at high risk. The risk of countries in West Asia and North Africa, such as Afghanistan, Iran, Israel, and Lebanon is also at high risk but slightly fluctuated from year to year. The research results show that the classification of national risks by media big data has strong consistency with existing national risk ratings, so this article proposes to use media big data to enhance the risk perception capabilities of countries along the “Belt and Road”.