A real-time risk assessment model for cross-border financial transactions based on big data technology

Abstract

The study applies the method of resampling to deal with unbalanced financial transaction data, which is resampled by the method of majority class weighted minority class oversampling. After data processing, the VaR-GARCH financial transaction risk assessment model is constructed. The financial transaction risk assessment method of this paper is compared with other risk assessment methods to test its assessment performance. Subsequently, taking the carbon financial market as an entry point, the trading price data of seven global carbon financial markets from 2021 to June 28, 2024, are selected for the study to assess the risk of the carbon transnational trading market in real-time. The risk assessment efficacy of this paper’s risk assessment model on both the AP and LC datasets has an overall advantage. Among the seven global carbon markets, the EU has the most drastic fluctuation in transaction prices, while the Chinese carbon market is the smoothest. The transaction price averages from highest to lowest are California-Quebec (85.59), South Korea (72.49), U.S. Regional Greenhouse Gas Emission Reduction Program (47.24), U.K. (44.80), China (37.26), New Zealand (34.35), and EU (34.34). California-Quebec had the highest average transaction price, while the EU had the lowest average transaction price. Transaction prices in China are the most stable, and log yield trends in the UK and South Korea are similar. The top three markets in terms of value-at-risk VaR are California-Quebec, South Korea, and the EU, and the smallest is the UK market.

Keywords

• resampling techniques
• garch-var model
• risk assessment
• cross-border financial transactions
• 68t42