Abstract:In decision-making systems, how to address uncertainty plays an important role for the improvement of system performance in uncertainty reasoning. Dempster–Shafer evidence (DSE) theory is an effective method to address uncertainty in decision-making problems by means of basic belief assignments (BBAs) and Dempster's combination rule. In the DSE theory, divergence measure between BBAs, which is beneficial for conflict information management in decision making, remains an open issue. In this paper, several generalized evidential divergences (EDs) are proposed and studied to measure the difference and discrepancy between BBAs in DSE theory, which have more universal applicability in decision theory. On this basis, a uniform $mathcal {BJS}$BJS divergence-based decision-making algorithm is devised to improve the decision level. Furthermore, the extensions of weighted $mathcal {BJS}$BJS to decision-making algorithms are discussed by considering not only subjective weights but also objective weights. Notably, this is the first work to propose the weighted $mathcal {BJS}$BJS divergence in DSE theory providing a promising way to analyze decision-making problems from different perspectives. Besides, experiments demonstrate the effectiveness and superiority of the proposed methods. Finally, the proposed $mathcal {BJS}$BJS-based decision-making algorithm is applied to pattern classification. The results validate that the proposed decision-making algorithm is beneficial for diverse real-world datasets and outperforms several well-known related works and demonstrates higher classification accuracy as well as robustness.

A Classification Tree Method Based on Belief Entropy for Evidential Data

Evidential Decision Tree Based on Belief Entropy

Belief Entropy Tree and Random Forest: Learning from Data with Continuous Attributes and Evidential Labels

Online active learning of decision trees with evidential data

An Improved Belief Entropy and Its Application in Decision-Making

Bagging likelihood-based belief decision trees

Learning Belief Regression Trees From Evidential Data

A Method of Inducing Ordered Basic Belief Assignment and Its Application in Risk Assessment of Tailing Dam

A Modified Belief Entropy in Dempster-Shafer Framework

Big Data Classification Using Belief Decision Trees: Application to Intrusion Detection

An Improved Belief Entropy in Evidence Theory

A new random forest method based on belief decision trees and its application in intention estimation

A new correlation belief function in Dempster-Shafer evidence theory and its application in classification

A less-greedy two-term Tsallis Entropy Information Metric approach for decision tree classification

DPT: An importance-based decision probability transformation method for uncertain belief in evidence theory

Generalized Divergence-based Decision Making Method with an Application to Pattern Classification

Transformation of basic probability assignments to probabilities based on a new entropy measure

Handwritten Digit Recognition Via Active Belief Decision Trees

A novel uncertainty modeling method in complex evidence theory for decision making

Uncertainty-Aware Fairness-Adaptive Classification Trees

Improving Decision Trees by Tsallis Entropy Information Metric Method.