Abstract Current research on financial distress prediction faces numerous unresolved issues. Specifically, financial market data often exhibits asymmetric volatility and heavy-tailed characteristics, leading to skewed distributions. Against this backdrop, while the Grey Wolf algorithm possesses powerful global search capabilities, it is prone to population imbalance when handling high-dimensional data. To address these problems, this paper proposes a quantum-inspired-optimized Grey Wolf algorithm (QC-GWO) for financial forecasting. Specifically, we propose a hybrid population initialization strategy. This strategy combines Latin hypercube sampling and Adaptive Tilt Gaussian Adversarial Learning (ASG-OBL) to improve population diversity and reduce sensitivity to local patterns. To further prevent premature convergence, we apply a quantum-inspired computing based t-distribution mutation strategy to update the global optimum. Furthermore, we introduce the concept of quantum quasi-inverse learning, continuously updating the quantum state to expand the search space, thereby improving convergence accuracy and reducing stagnation. The experimental results conducted on the dataset of listed companies show that the proposed quantum computing optimized grey wolf algorithm (QC-GWO) outperforms the traditional grey wolf algorithm and other advanced optimization algorithms in terms of accuracy, recall, and AUC. The financial crisis prediction model based on the grey wolf algorithm and optimized by quantum-inspired is superior to the current optimization algorithms in terms of prediction accuracy, providing a new practical tool for the early detection of financial distress in listed companies.
Wen et al. (Tue,) studied this question.