Enhanced Risk Detection in Digital Finance via Improved Fish Swarm Optimization Enhanced Back Propagation Neural Network (IFSO-BPNN)

Abstract

As early warning system for digital finance hazards increased, this study considersboth macro and market factors. By providing a neural network prognosis of undesirable events, early warning systems can aid in averting commercial and economic catastrophes. In several industries, including banking, artificial intelligence (AI) has advanced quickly. This study proposed an innovative model called Improved Fish Swarm Optimization based on Back Propagation Neural Network (IFSO-BPCNN). An overview of primaryalerttechnology, their evolution along with their application in many contexts—particularly in the financial and economic sectors and how the proposed model used in risk assessment are presented in this work. It looks into how primaryalert systems might be incorporated to forecast and identify unfavorable actions, particularly in the corporate, financial as well as economic sectors. The data quality applied structured preprocessing and normalization using a real-world financial dataset that includes credit scores, transactional behavior, and user activity patterns from digital platforms. The proposed IFSO algorithm optimizes the weight initialization and learning trajectory of the BPNN by dynamically modifying search behavior parameters. By offering a digital finance study of the creation and application of primary alert technology for social and economic growth, the paper adds to the body of current literature. Data from 2012 to 2024 were used to evaluate the system's effectiveness, allowing for dynamic early warning assessments of digital financial threats. The scores having an exactness of 98.7, 97.8 as well as F1 score of 95.7, the IFSO-BPNN model outperformed four comparable models in terms of performance metrics.