Résumé: Due to an increase in decision making sophistication and risk, behavioural three-way decision making has replaced traditional two-way decision making. Accordingly, this study proposes a novus three-way multi-attribute decision making model by creating three-way decision rules in context of multi-attribute decision-making with two key behavioural theories, namely, regret and prospect theories. To capture uncertainty in the evaluation information, q-rung orthopair fuzzy set is implemented in the information system. After that, a new method is forwarded based on correlation coefficient and variance to calculate the weights of attributes. In light of the conventional Mahalanobis distance, q-rung orthopair fuzzy Mahalanobis distance is developed. Then, regret-based dominated relation and rejoice-based dominating relation are defined. Thereafter, a new procedure is developed to estimate the conditional probability according to two above-mentioned relations. Three kinds of scoring functions are developed with the help of prospect theory, based on which, despondent, expectant, and balanced tactics are developed to illustrate three mindsets of decision maker. Ultimately, three sizeable datasets gathered from the KEEL and UCI databases are used to demonstrate the legitimacy, resiliency, and supremacy of the constructed model.

Langue: Anglais