Publication

Abstract : This paper explores the effectiveness of different similarity measures for characterizing vertices and edges in rough graphs, which were introduced to handle imprecise and uncertain information. The authors examine traditional similarity measures like the Jaccard index, Dice coefficient, and overlap measure in this context. Additionally, a new ζ-labeling similarity measure for rough graphs is proposed. The main goal is to perform a comparative analysis evaluating the performance of these diverse similarity measures when applied to rough graphs. Furthermore, the paper computes the energy of rough graphs, defined as the sum of absolute eigenvalues, to demonstrate the superior potency of the proposed ζ-labeling measure compared to the other similarity measures considered. Overall, this work aims to advance techniques for assessing similarity in rough graphs, which have applications in dealing with vague and imprecise data.