Three-way multi-criteria decision-making with linguistic information using prospect-regret theory and generalised TODIM for heart failure data analysis.

Cardiovascular disease is a leading global cause of mortality, and timely monitoring can help assess its severity. Electronic medical records contain information on patients’ symptoms, physical characteristics, and clinical laboratory test results. These data can be used in biostatistical analysis to identify patterns and relationships that may not be apparent to medical practitioners. However, when experts confront risk in complex settings, their bounded rationality may influence their decision-making behaviour. Moreover, the irrational behaviour of a single expert can substantially affect decision outcomes. Motivated by these observations, this paper develops a novel three-way multi-criteria decision-making technique in the linguistic Z-number (LZN) environment based on prospect and regret theories (PRT), abbreviated as 3WMCPRT. The proposed model can predict how long a patient is likely to survive on the basis of clinical records and can also identify the most significant features in those records. In the 3WMCPRT model, fuzzy decision objects (FDOs), conditional probabilities (CPs), and threshold values are the three key components used to divide alternatives into three regions and rank them. To compute these components, the generalised TODIM-based FDOs manage experts’ personal regrets, the PRT-based CPs capture the influence of experts’ rational and irrational behaviour on decision-making, and the threshold values are derived from relative loss functions. Comparative analyses with alternative methodologies demonstrate the feasibility, stability, and superiority of the proposed 3WMCPRT model.