Investigation on Market Manipulation of Digital Currency Based on Artificial Intelligence Technology

Abstract

As blockchain technology drives the global expansion of the digital currency market, the widespread adoption of high-frequency trading and cross-market arbitrage strategies poses dual challenges to traditional regulatory measures in terms of timeliness and accuracy. This study constructs a hybrid neural network model that integrates supervised and unsupervised learning to explore multi-dimensional feature fusion paths between on-chain data from blockchain and secondary market price data. Based on dynamic game theory, an intelligent regulatory sandbox system is designed, incorporating on-chain address reputation scoring mechanisms and liquidity smart contract circuit breakers to achieve real-time warnings and responses to market manipulation behaviors. Furthermore,a distributed regulatory framework built on zero-knowledge proof technology is proposed, providing a feasible solution for establishing a penetrating regulatory system while ensuring transaction privacy.