Cascade Kronecker neuro–fuzzy network based influential node identification

The rapid growth of information technology has led to increasingly complex networks, making the spread of rumors and diseases more challenging to manage. Identifying influential nodes efficiently and accurately is crucial for predicting and controlling network dynamics. Existing methods often fail to account for the impact of edges or rely on global network configurations, leading to inaccuracies and high computational costs. Here, Cascade Kronecker Neuro-Fuzzy Network (KNFN) is introduced for influential node identification. At first, initial influential node identification is accomplished based on chord distance based betweenness, Betweenness Centrality (BC), closeness of node and Degree Centrality (DC). Then, community creation is carried out utilizing chord distance. In addition, features for the community are considered and based on considered features community detection is done. The detection of community is executed employing Cascade KNFN, which is an assimilation of the Deep Kronecker Network (DKN) with Cascade Neuro-Fuzzy Network (NFN). Lastly, the right community to merge is identified by exploiting chord distance. The evaluation is conducted based on chord distance based betweenness and communication cost and the Cascade KNFN acquired chord distance based betweenness of 0.624 and communication cost of 1.046. This study provides a robust framework for managing complex networks and improving prediction and control in dynamic systems. • The community creation is carried out for community detection utilizing chord distance. • Based on considered features for community detection is conducted by hybrid network. • After detection, right community for merging is determined exploiting chord distance.