Influence maximization algorithm based on community structure and applicable to paper cooperation network

Abstract

The invention provides an influence maximization algorithm based on a community structure and applicable to a paper cooperation network. The influence maximization algorithm comprises the following steps of:1) in a community discovering phase, a, constructing the paper cooperation network, b, merging local communities, c, constructing a new network image, and d, ending; and 2) in a seed node selecting phase, a, calculating the influence of each community, b, selecting the corresponding nodes in the community largest in influence, and c, ending. According to the invention, a new solution scheme is provided for the influence maximization problem of the paper cooperation network, and indicated by results, for the ICM model, the COMAX algorithm provided by the invention is close to a greedy algorithm in influence coverage range, and the time efficiency is very good.

Description

technical field [0001] The invention relates to a method for solving the problem of maximizing the influence of a paper cooperation network, in particular to a method for solving the problem of maximizing the influence based on a community structure. Background technique [0002] In recent years, online social networks have developed rapidly, and more and more social networking sites have emerged. The information dissemination in these social networks has surpassed real life both in terms of scale and efficiency. The influence maximization problem focuses on how to select a fixed number of seed nodes to maximize the coverage of information dissemination. When we need to do research or in-depth understanding of a certain subject or field, we will not look at all the information in this field, we will select some works of high-impact authors, how to find these high-impact authors is the selection of seed nodes the process of. [0003] In 2003, Kempe, Kleinberg and Tardos [M...

AI Technical Summary

Problems solved by technology

However, the time complexity of the greedy algorithm is very high. It does not consider the degree and division of the network, and does not consider the community structure of the network. Every time a seed node is selected, the influence of each seed node needs to be recalculated, and the time efficiency is relatively low.

[0004] In 2007, in response to the high time complexity of the greedy algorithm, Leskovec et al [Cost-effectiveOutbreak Detection in Networks] used the sub-model characteristics in impact maximization to propose the optimization strategy of "LazyForward", and proposed the CELF algorithm, CLEF Due to the use of submodel characteristics, the algorithm reduces the calculation scale in the seed selection stage and improves the efficiency of the greedy algorithm to a certain extent, but it is not suitable for large-scale social networks.

[0006] Many influence maximization algorithms do not take into account the community structure of the network, but the connections between nodes within the community are closer than those outside the community. Correspondingly, in the process of information dissemination, the possibility of nodes activating other nodes in the same community is also higher. It is more likely to activate the external nodes of the community

Images