CPR-TOPSIS: A novel algorithm for finding influential nodes in complex networks based on communication probability and relative entropy

Abstract

How to evaluate the importance of nodes and identify influential nodes in complex networks is a very significant research in the field of network science. Most of existing algorithms neglect the relationship between a node and its neighbors to evaluate the importance of nodes in networks. In this work, we first define nodes communication probability sequence by making use of the length and number of shortest paths between node pairs. Then the traditional binary adjacency matrix is converted into correlation matrix through relative entropy. Based on information Communication Probability and Relative entropy (CPR), an improved Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), called CPR-TOPSIS, is presented for identifying influential nodes in complex networks from the view of global, local and location information dimensions. The proposed algorithm has been compared with eight state-of-the-art algorithms on several real-world networks to verify the performance. Experimental results show that CPR-TOPSIS has better performance in terms of monotonicity, resolution, ranking accuracy, imprecision function and top-10 nodes.