Managing Overconfidence Behaviors From Heterogeneous Preference Relations in Linguistic Group Decision Making

Abstract

The linguistic information of decision makers (DMs) often implies individual preferences and behavioral traits. Reaching consensus in linguistic group decision making (GDM) requires understanding the overconfidence behaviors exhibited by DMs and implementing effective management. However, owing to the differences in individual expression and understanding, the management of the overconfidence behaviors of DMs must effectively handle heterogeneous preference information and personalized individual semantics (PIS). To solve this problem, this study manages overconfidence behaviors exhibited by DMs in linguistic GDM by considering PIS and multiple self-confidence levels. Specifically, DMs utilize flexible linguistic representation models to provide preference values over alternatives and corresponding self-confidence levels, thereby generating heterogeneous preference relations with self-confidence (PRs-SC). We then integrate heterogeneous preference relations into unified linguistic distribution expressions. In addition, we transform linguistic PRs-SC into additive PRs-SC using a PIS model. We detect and manage individuals' overconfidence behaviors to decrease the negatively impact of overconfidence on consensus efficiency and the decision quality. Subsequently, DMs' preference values and self-confidence levels are utilized to generate modification suggestions to achieve a consensus. Finally, we use a loan selection on a peer-to-peer lending platform as an example to illustrate and demonstrate the usability, effectiveness, and feasibility of our model through a simulation and comparative analysis.