Collaborative Prediction Algorithm for Investment and Financing Network System Based on Improved Hypergraph Convolution and Ant Colony Algorithm

After more than 20 years of rapid development, the Chinese financial market is experiencing a decrease in user traffic dividends. The overall industry growth rate is slowing down, yet the market’s potential demand remains substantial. The industry is transitioning from a phase focused on incremental growth to one that equally emphasizes incremental and existing stock growth, accelerating the exploration of new development models. This paper proposes the CPA-FHA, a collaborative prediction algorithm for investment and financing networks based on improved hypergraph convolution and ant colony algorithms, by integrating TransD embedding method, hypergraph convolution, spatial feature extraction, and ant colony algorithms. The proposed method is compared and analyzed using real-world investment and financing network datasets from Chinese listed companies, Chinese financial institutions, and American listed companies. It is found that, compared to classical algorithms, the Collaborative Prediction Algorithm for Financing and Investment Network System - Forecast and Hazard Analysis (CPA-FHA) has higher accuracy and efficiency in predicting investment and financing network systems.