Open Access

ARTICLE

Identifying Influential Communities Using IID for a Multilayer Networks

by C. Suganthini^*, R. Baskaran

Department of Computer Science and Engineering College of Engineering, Guindy, Anna University, Chennai, Tamilnadu, India

* Corresponding Author: C. Suganthini. Email:

Intelligent Automation & Soft Computing 2023, 36(2), 1715-1731. https://doi.org/10.32604/iasc.2023.034019

Received 04 July 2022; Accepted 19 August 2022; Issue published 05 January 2023

Abstract

In online social networks (OSN), they generate several specific user activities daily, corresponding to the billions of data points shared. However, although users exhibit significant interest in social media, they are uninterested in the content, discussions, or opinions available on certain sites. Therefore, this study aims to identify influential communities and understand user behavior across networks in the information diffusion process. Social media platforms, such as Facebook and Twitter, extract data to analyze the information diffusion process, based on which they cascade information among the individuals in the network. Therefore, this study proposes an influential information diffusion model that identifies influential communities across these two social media sites. Moreover, it addresses site migration by visualizing a set of overlapping communities using hyper-edge detection. Thus, the overlapping community structure is used to identify similar communities with identical user interests. Furthermore, the community structure helps in determining the node activation and user influence from the information cascade model. Finally, the Fraction of Intra/Inter-Layer (FIL) diffusion score is used to evaluate the efficiency of the influential information diffusion model by analyzing the trending influential communities in a multilayer network. However, from the experimental result, it observes that the FIL diffusion score for the proposed method achieves better results in terms of accuracy, precision, recall and efficiency of community detection than the existing methods.

---

Keywords

---