Complex Network-based Analysis On Propagation Mechanism Of Public Sentiment Information

In recent years,with the rapid development of information technology and the widespread use of intelligent communication tools,the online social networks have become an important way in sharing and spreading information.As a typical representative of Web 2.0era,Sina microblog has become one of the ten largest social networking sites in the world after 8 years of development,and has surpassed Twitter to become the largest independent social media company in the world.Hundreds of millions of users on microblog have formed a new network ecosystem and a huge information communication network.It can be said that it brings a upgrading of the ideological system,not only changing the way of people's life and production,but also promoting the innovation of society and business forms.The purpose of this paper is to reveal the information propagation mechanisms and the inherent laws both microcosmically and macroscopically according to dynamic model and complex network theory.Theoretically,the research achievements enrich the methods of public opinion dissemination theory,promoting inter-disciplinary studies and advance the innovation of data mining model in the era of big data.Moreover,from the part of practical level,the research on the internal mechanism of public opinion dissemination provides solutions for the emergency events and the research on the evolution process of public opinion transmission provides an effective way to absorb the opinions of the public and provides a reliable basis for the "precise marketing" of enterprise investors.(1)From the perspective of single layer network,the cooperative and competitive dynamics model for information propagation is proposed.We introduce the analytical framework of population dynamics,and propose an information propagation model based on velocity.And then,we propose a Lotka-Volterra cooperative and competition propagation model,thus to expounding the mechanism of information network and working out its system equilibrium point and its stability.A simulation analysis is given according to the actual data from Sina Weibo.(2)From the perspective of coupling network,we construct the coevolution model for public opinion information in the double-layered coupling network,uses Price network model and WS network model to simulate on-line/off-line network layers respectively,establishes one-to-one correspondence between the nodes by means of assortative connection between the layers and analyzes the on-line/off-line coevolution process of public opinion information on this basis.(3)From the perspective of user behavior,we find the most influential user nodes.This paper considerates the relationship between user and social network,and proposes MPSO algorithm which applies island model to the improvement of standard particle swarm algorithm.The algorithm considerates the information from user attribute and social network factors,so it can overcome the influence of artificial followers.At the same time,MPSOalgorithm increases the diversity of particles and avoids to fall into local optimum.Then,we compare the Behavior-Relationship Rank algorithm,Page Rank algorithm with MPSO algorithm to prove the accuracy and reliability of it.(4)Based on the complexity of information dissemination network and the original of bad information,we construct overall supervising strategy framework including technology and system according to the national conditions.