Preference and attitude in parameterized knowledge measure for decision making under uncertainty

Kaihong Guo,Hao Xu
DOI: https://doi.org/10.1007/s10489-021-02317-2
IF: 5.3
2021-03-11
Applied Intelligence
Abstract:We develop in this paper a novel entropy-independent knowledge measure (KM), with which to reveal some significant aspects of psychological cognition hidden in the handling of intuitionistic fuzzy sets (IFSs). We briefly discuss the two facets of knowledge associated with an IFS, i.e., the information content and the information clarity. We then establish, based on the latest axiomatic definition of KM in the context of IFSs, a novel parameterized KM in which two significant aspects of psychological cognition are considered, personal attitude and preference to be exact. We believe that the KM provided in this manner could truly capture the unique features of an IFS, including the potential knowledge related to the specificity and non-specificity of an IFS, the amount of which depends actually on users' character traits. The developed KM is equipped with two parameters, one of which expresses the type of attitude towards the non-specificity of an IFS while the other indicates the degree of personal preference between those two facets of knowledge. We also show that some existing measures can be obtained as particular cases of this general model. Finally, we illustrate the application of this measure in decision making under uncertainty.
computer science, artificial intelligence
What problem does this paper attempt to address?
The paper attempts to address the issue of how to more effectively measure the amount of knowledge in Intuitionistic Fuzzy Sets (IFSs) in uncertain decision-making. Specifically, the paper proposes a new entropy-independent Knowledge Measure (KM) aimed at revealing some important aspects of psychological cognition when dealing with Intuitionistic Fuzzy Sets, particularly the impact of personal attitudes and preferences on the knowledge measure. ### Main Issues: 1. **Limitations of Existing Knowledge Measures**: Existing knowledge measures mainly rely on entropy, but entropy cannot distinguish certain Intuitionistic Fuzzy Sets with the same uncertainty, especially when different personalities and character traits need to be considered. 2. **Consideration of Psychological Cognition**: Existing knowledge measures do not adequately consider the impact of personal attitudes and preferences on the knowledge measure, which is an important factor in practical applications. 3. **Flexibility and Personalization**: Existing knowledge measures lack flexibility and cannot flexibly capture the potential knowledge related to the specificity and non-specificity of Intuitionistic Fuzzy Sets according to the user's character traits. ### Solution: The paper proposes a new Parameterized Knowledge Measure (PKM), which includes two parameters: - **α**: Represents the type of attitude towards the non-specificity of Intuitionistic Fuzzy Sets. - **β**: Represents the degree of preference between information content and information clarity. By introducing these two parameters, the new knowledge measure can more flexibly capture the potential knowledge related to the specificity and non-specificity of Intuitionistic Fuzzy Sets and reflect the different character traits and preferences of users. ### Application Prospects: This new knowledge measure has broad application prospects in uncertain decision-making, especially in complex situations where personal attitudes and preferences need to be considered, such as image processing, decision support systems, etc.