A Hybrid AI Framework Combining Causal Nonmonotonic Reasoning with VMD-GNN and IGWO-DBN for Decision-Making under Big Data

Abstract

With the deep integration of big data and Artificial Intelligence, data-driven predictive models face key challenges such as missing causality and rigid reasoning. Traditional approaches rely on statistical correlations, which struggle to handle uncertainty, conflicting evidence, and sparse data in dynamic environments, resulting in poor interpretability of decision-making. This paper proposes an Artificial Intelligence decision-making model enhanced by causal nonmonotonic reasoning. This model combines Variational Mode Decomposition and Graph Neural Network to process dynamic temporal data and graph structures, and introduces an improved Grey Wolf Optimizer to optimize the parameters of Deep Belief Network, thus constructing a decision framework that integrates dynamic causal discovery, revocable rule inference, and counterfactual evaluation. The experimental results on the MIMIC-III medical dataset and IEEE-CIS financial dataset show that the proposed model achieves an average processing error as low as 0.02, which is superior to the three baseline models Graph Attention Network- Recurrent Neural Network, Particle Swarm Optimization-Ant Colony Optimization, and Genetic Algorithm-Long Short-Term Memory Network. In financial volatility prediction, the model achieved a root mean square error of 0.12, a coefficient of determination of 0.92, and a peak-valley capture rate of 0.89. It also maintained a robust recall rate ranging from 0.8 to 0.98 across confidence intervals of 0 to 0.6, demonstrating the best overall performance. These results indicate that the proposed model shows strong accuracy and adaptability in complex decision-making scenarios. It effectively addresses the problems of causal blindness and rigid reasoning in current Artificial Intelligence systems, offering a new approach to trustworthy Artificial Intelligence and promoting intelligent and efficient development in key fields such as healthcare and finance.