Sensitivity Analysis for Fuzzy Similarity Measures

• Published in: ASM Science Journal
• Main Author: Rasmani K.A.; Shahari N.; Mohd Razif S.N.R.; Muhamad Zambri F.N.; Razali N.F.; Wie S.S.; Rusli A.D.; Rosli N.N.
• Format: Article
• Language: English
• Published: Akademi Sains Malaysia 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85200338942&doi=10.32802%2fASMSCJ.2023.1180&partnerID=40&md5=c41dd3dd40ff8c4fd7f5718cc35d319a

Fuzzy similarity measure (FSM) is the method used to calculate similarity between fuzzy sets. Various techniques have been formulated, but it can be observed that there is a lack of formal method to determine which method is better or suitable to be used for certain applications. Many researchers make comparisons between methods based on selected cases only, which is, in essence not enough to conclude which method is better. This study proposes sensitivity analysis where parameter adjustment will be made relative to the height, distance, area, or perimeter so that the similarity values obtained can be compared and analysed. The result shows that the parameter adjustment will result in changes of similarity values. Therefore, the sensitivity analysis should be regarded as very important when comparing between two fuzzy numbers. Hence, the proposed method is expected to be very useful in assisting researchers to determine the behaviour of the selected fuzzy similarity measure. This finding suggests that the current practice by comparing fuzzy numbers based on selected sample are not enough to provide conclusive result. The concept of analysis introduced in this study can be a starting point for more systematic analysis on fuzzy similarity measures. This hopefully will open for broader implementations of fuzzy similarity measures in real world decision-making. © (2024), (Akademi Sains Malaysia). All Rights Reserved.