Information measures based on similarity under neutrosophic fuzzy environment and multi-criteria decision problems

Abstract

The neutrosophic fuzzy set (NF-set) is a unique hybrid structure that combines the essence of the fuzzy and neutrosophic sets. It is an effective mathematical tool for dealing with indeterminate and inconsistent information in situations where the data are imprecise or vague. This paper defines the concepts of similarity-based information measures, including entropy and cross-entropy, between NF-sets. It is the first time these concepts have been introduced since the NF-sets were defined. We provide their interesting properties through well-proven theorems. In addition, we also propose a novel and efficient algorithm to make multi-criteria decision(s) using these information measures with a clear step-by-step illustration. It overcomes the limitations of the original algorithm when evaluating the criteria in two aspects: qualitative and quantitative. A real-world experiment is then conducted to assist students in choosing the optimal subject group(s) for the Vietnamese national high school graduation examination. Experimental results show that our method outperforms the previous original method when giving recommendations with sixteen correct cases, three acceptable cases, and one noisy case out of twenty real-life cases. Finally, the experimental results are presented visually, analyzed rigorously, and discussed carefully, intending to verify the validity and feasibility of the proposed algorithm. The Python source code for experiments is publicly available on Github.