Inter Observer Variability Assessment through Ordered Weighted Belief Divergence Measure in MAGDM Application to the Ensemble Classifier Feature Fusion

TL;DR

The work addresses MAGDM under uncertainty and expert conflict by introducing an Evidential MAGDM framework that combines basic probability assignments, ordered weighted belief/plausibility, and a generalized divergence measure to quantify inter-expert variability. The methodology yields a complete workflow for generating BPAs, computing weighted beliefs and plausibilities, and fusing expert inputs to obtain final rankings. It includes an illustrative numerical example and a real-world retinal diagnosis application, where an ensemble of multi-scale OCT features fused via the Evidential MAGDM approach achieves superior performance (e.g., accuracy about 0.911) compared with baselines. The results demonstrate the method's ability to handle imprecision and conflicts without relying on fixed expert weights, highlighting its potential for robust decision support in medical imaging and similar MAGDM contexts.

Abstract

A large number of multi-attribute group decisionmaking (MAGDM) have been widely introduced to obtain consensus results. However, most of the methodologies ignore the conflict among the experts opinions and only consider equal or variable priorities of them. Therefore, this study aims to propose an Evidential MAGDM method by assessing the inter-observational variability and handling uncertainty that emerges between the experts. The proposed framework has fourfold contributions. First, the basic probability assignment (BPA) generation method is introduced to consider the inherent characteristics of each alternative by computing the degree of belief. Second, the ordered weighted belief and plausibility measure is constructed to capture the overall intrinsic information of the alternative by assessing the inter-observational variability and addressing the conflicts emerging between the group of experts. An ordered weighted belief divergence measure is constructed to acquire the weighted support for each group of experts to obtain the final preference relationship. Finally, we have shown an illustrative example of the proposed Evidential MAGDM framework. Further, we have analyzed the interpretation of Evidential MAGDM in the real-world application for ensemble classifier feature fusion to diagnose retinal disorders using optical coherence tomography images.

Figures (5)

• Figure 1: Linguistic values membership function with $(H+1)$ terms
• Figure 2: The proposed framework for group decision-making by analyzing the inter-observational variability between each alternative corresponding to the attributes for handling the impreciseness and conflicts between the group of experts.
• Figure 3: The proposed Evidential MAGDM feature fusion module integrated into the ensemble classifier to analyze inter-observational variability between the features.
• Figure 4: Description of various retinal disorders includes in the OCTDL dataset.
• Figure 5: The confusion matrix of the proposed Evidential MAGDM method in the homogeneous ensemble classifier feature fusion with different weights (a) EfficientB0 (Evidential MAGDM), (b) EfficientB0$_{(1,0,0)}$, (c) EfficientB0$_{(0,1,0)}$, (d) EfficientB0$_{(0,0,1)}$, (e) EfficientB0$_{(0.5,0.5,0)}$, (f) EfficientB0$_{(0.5,0.25,0.25)}$.

Theorems & Definitions (12)

• Definition 1
• Definition 2
• Definition 3
• Definition 4
• Definition 5
• Definition 6
• Definition 7
• Definition 8
• Definition 9
• Definition 10