The precise selection of key flight parameters is fundamental to enhancing aircraft condition monitoring and risk warning capabilities. However, existing methods typically rely on a single source of information, i.e., either solely expert judgments or solely objective flight data, and lack effective mechanisms to reconcile conflicts between subjective opinions and objective data characteristics, which limits their applicability in complex aviation safety scenarios. To address this issue, a flight parameter selection method based on dual-domain rough sets and three-way decision theory is proposed in this paper. First, regret theory is introduced to quantify experts’ psychological preferences, and a subjective evaluation model integrating both psychological and absolute agreement is constructed. Second, a subjective–objective conflict information system is established within a dual-domain framework. Based on this system, bidirectional decision rules are designed to simultaneously consider positive-domain and negative-domain conditional probabilities, through which candidate sets of key flight parameters are generated. Finally, a new Bayesian minimum loss criterion is designed to determine the optimal parameter set. Experimental results demonstrate that the accuracy and robustness of flight parameter selection are improved by the proposed method while interpretability is maintained, offering reliable decision support for aviation safety analysis.
Yan et al. (2026) studied this question.