NPNT: Non-oscillating Process Negation Transformation of mass functions and a negation-based discounting method in information fusion

Qianli Zhou, Yong Deng

Abstract: Yager proposes a Probability Negation Transformation (PNT) based on the maximum entropy distribution. After continuous PNTs, the Probability Mass Function (PMF) reaches the maximum entropy distribution, i.e., an uniform distribution. In this paper, we propose a new negation transformation method based on Belief Evolution Network (BEN) and Disjunctive Combination Rule (DCR), called Non-oscillating Process Negation Transformation (NPNT). As a cognitive negation transformation, NPNT expresses the negation from known to unknown. The Full Causality Probability Transformation (FCPT) is used to connect the PMF and Basic Belief Assignment (BBA) under NPNT. Based on the proposed method, we extend the Shafer’s discounting method from the perspective of DCR, which realizes less commitment under a same discount coefficient.