@Article{iasc.2022.024561, AUTHOR = {Debasis Mohapatra, Sourav Kumar Bhoi, Kalyan Kumar Jena, Chittaranjan Mallick, Kshira Sagar Sahoo, N. Z. Jhanjhi, Mehedi Masud}, TITLE = {Core-based Approach to Measure Pairwise Layer Similarity in Multiplex Network}, JOURNAL = {Intelligent Automation \& Soft Computing}, VOLUME = {34}, YEAR = {2022}, NUMBER = {1}, PAGES = {51--64}, URL = {http://www.techscience.com/iasc/v34n1/47352}, ISSN = {2326-005X}, ABSTRACT = {Most of the recent works on network science are focused on investigating various interactions among a set of entities present in a system that can be represented by multiplex network. Each type of relationship is treated as a layer of multiplex network. Some of the recent works on multiplex networks are focused on deriving layer similarity from node similarity where node similarity is evaluated using neighborhood similarity measures like cosine similarity and Jaccard similarity. But this type of analysis lacks in finding the set of nodes having the same influence in both the network. The discovery of influence similarity between the layers of multiplex networks helps in strategizing cascade effect, influence maximization, network controllability, etc. Towards this end, this paper proposes a pairwise similarity evaluation of layers based on a set of common core nodes of the layers. It considers the number of nodes present in the common core set, the average clustering coefficient of the common core set, and fractional influence capacity of the common core set to quantify layer similarity. The experiment is carried out on three real multiplex networks. As the proposed notion of similarity uses a different aspect of layer similarity than the existing one, a low positive correlation (close to non-correlation) is found between the proposed and existing approach of layer similarity. The result demonstrates that the degree of coreness difference is less for the datasets in the proposed method than the existing one. The existing method reports the coreness difference to be 40% and 18.4% for the datasets CS-AARHUS and EU-AIR TRANSPORTATION MULTIPLEX respectively whereas it is found to be 20% and 8.1% using proposed approach.}, DOI = {10.32604/iasc.2022.024561} }