复杂网络中异质社会传播动力学研究
发布时间：2018-08-11 18:37
【摘要】：近年来,随着科学技术的飞速发展,人类进入网络时代,社会传播现象随处可见。各类软件在各大应用平台宣传下载、各种时尚元素在大街小巷随处可见、各种谣言舆论在各大网站广泛散播等传播行为,在提高了人们生活水平的同时,也存在危害人类利益的安全隐患。因此,研究社会传播机制并制定相应策略对传播进行控制是当今重大研究课题之一。目前,社会传播过程中个体采纳行为的异质性、网络结构的异构性、个体采纳行为的记忆性、传播中的非冗余传播和非完全邻居传播等综合因素对社会传播的综合影响尚未被深入研究,故本文基于复杂网络理论,进行一系列社会传播动力学研究和免疫研究,并取得一定成果:1.提出异质社会传播模型并进行理论解析。将社会传播过程中非冗余传播、非马尔科夫特性、非完全邻居传播、采纳阈值的异质性、网络不同结构特性等因素综合考虑(统称为异质性),提出异质社会传播模型;同时对边划分理论进行改进,提出异质边划分理论,进行稳态分析、爆发阈值解析,以及为后续实验仿真进行理论拟合;2.进行大量异质社会传播实验仿真和理论拟合。为了更深刻理解异质性对社会传播的影响,本文在度度无关联的ER网络和SF网络中,基于异质社会传播模型,通过大量的实验仿真社会传播过程并进行理论拟合。实验发现:当采纳阈值与度完全正关联时,采纳阈值的异质性越强,行为传播先被抑制后被促进;完全负关联时,采纳阈值的异质性越强,越有利于传播。当部分正关联时,采纳阈值与度越不相关,越促进传播;当部分负关联时,采纳阈值与度越不相关,越抑制传播。其中最终采纳比例随网络传播率增加,会出现一级相变、二级相变和混合相变现象。3.对异质社会传播进行免疫研究。基于异质社会传播模型,分别使用度中心性、接近中心性和k-core中心性免疫节点,分别在ER网络和SF网络上进行异质社会传播,观察对传播范围的影响并进行免疫性能对比。实验发现:在ER网络和SF网络上,当采纳阈值与度无关联和完全负关联时,均建议选用度中心性免疫;当采纳阈值与度完全正关联时,在采纳阈值的异质性较弱时,均适合选用度中心性免疫,但在采纳阈值的异质性较强时,ER网络适合选用k-core中心性免疫,SF网络上,各中心性免疫无法抑制传播。
[Abstract]:In recent years, with the rapid development of science and technology, the phenomenon of social communication can be seen everywhere. Various kinds of software are propagandized and downloaded in various application platforms, all kinds of fashion elements can be seen everywhere in the streets and alleyways, and various kinds of rumors and public opinions are widely disseminated in the major websites, so as to improve people's living standards at the same time. There are also hidden dangers that endanger the interests of mankind. Therefore, it is one of the important research topics to study the mechanism of social communication and formulate corresponding strategies to control it. At present, the heterogeneity of individual adoption behavior, the heterogeneity of network structure, the memory of individual adoption behavior in the process of social communication, The comprehensive influence of non-redundant propagation and incomplete neighbor communication on social communication has not been deeply studied. Therefore, based on the theory of complex network, this paper carries out a series of studies on social communication dynamics and immunity. And achieved some results: 1. The model of heterogeneous social communication is put forward and analyzed theoretically. Considering the factors of non-redundant communication, non-Markov characteristic, incomplete neighbor communication, adopting the heterogeneity of threshold value and different structure of network, the heterogeneous social communication model is put forward. At the same time, the theory of edge partition is improved, the theory of heterogeneity edge partition is put forward, the steady-state analysis is carried out, the threshold of explosion is analyzed, and the theoretical fitting for the subsequent experiment simulation is carried out. A large number of heterogeneous social transmission experiments and theoretical simulation. In order to better understand the influence of heterogeneity on social communication, this paper simulates the process of social communication through a large number of experiments and carries out theoretical fitting in the unrelated ER and SF networks, based on the heterogeneous social communication model. It is found that when the adoption threshold is positively correlated with the degree, the stronger the heterogeneity of the adoption threshold is, the more the behavior propagation is suppressed and then promoted, and the stronger the heterogeneity of the adoption threshold is when the adoption threshold is completely negative correlation. When the partial positive correlation is positive, the adoption threshold is more irrelevant to the degree, the more the propagation is promoted, the more irrelevant the adoption threshold is to the degree of partial negative correlation. Among them, the final adoption ratio increases with the network propagation rate, there will be first order phase transition, second order phase transition and mixed phase transition phenomenon. 3. The immune study on the transmission of heterogeneous society is carried out. Based on the heterogeneous social transmission model, the heterogeneous social transmission was carried out on ER network and SF network, respectively, using degree centrality, proximity and k-core central immune nodes. The effects on the transmission range were observed and the immune performance was compared. It is found that in ER and SF networks, when the adoption threshold is not correlated with degree and is completely negative, it is recommended to choose the degree central immunity, and when the adoption threshold and degree are completely positively correlated, the heterogeneity of the adoption threshold is weak. All of them are suitable for the selection of degree central immunity, but when the heterogeneity of threshold is strong, ER network is suitable for selecting k-core central immune SF network, each central immunity can not inhibit the transmission.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O157.5
本文编号：2177872
[Abstract]:In recent years, with the rapid development of science and technology, the phenomenon of social communication can be seen everywhere. Various kinds of software are propagandized and downloaded in various application platforms, all kinds of fashion elements can be seen everywhere in the streets and alleyways, and various kinds of rumors and public opinions are widely disseminated in the major websites, so as to improve people's living standards at the same time. There are also hidden dangers that endanger the interests of mankind. Therefore, it is one of the important research topics to study the mechanism of social communication and formulate corresponding strategies to control it. At present, the heterogeneity of individual adoption behavior, the heterogeneity of network structure, the memory of individual adoption behavior in the process of social communication, The comprehensive influence of non-redundant propagation and incomplete neighbor communication on social communication has not been deeply studied. Therefore, based on the theory of complex network, this paper carries out a series of studies on social communication dynamics and immunity. And achieved some results: 1. The model of heterogeneous social communication is put forward and analyzed theoretically. Considering the factors of non-redundant communication, non-Markov characteristic, incomplete neighbor communication, adopting the heterogeneity of threshold value and different structure of network, the heterogeneous social communication model is put forward. At the same time, the theory of edge partition is improved, the theory of heterogeneity edge partition is put forward, the steady-state analysis is carried out, the threshold of explosion is analyzed, and the theoretical fitting for the subsequent experiment simulation is carried out. A large number of heterogeneous social transmission experiments and theoretical simulation. In order to better understand the influence of heterogeneity on social communication, this paper simulates the process of social communication through a large number of experiments and carries out theoretical fitting in the unrelated ER and SF networks, based on the heterogeneous social communication model. It is found that when the adoption threshold is positively correlated with the degree, the stronger the heterogeneity of the adoption threshold is, the more the behavior propagation is suppressed and then promoted, and the stronger the heterogeneity of the adoption threshold is when the adoption threshold is completely negative correlation. When the partial positive correlation is positive, the adoption threshold is more irrelevant to the degree, the more the propagation is promoted, the more irrelevant the adoption threshold is to the degree of partial negative correlation. Among them, the final adoption ratio increases with the network propagation rate, there will be first order phase transition, second order phase transition and mixed phase transition phenomenon. 3. The immune study on the transmission of heterogeneous society is carried out. Based on the heterogeneous social transmission model, the heterogeneous social transmission was carried out on ER network and SF network, respectively, using degree centrality, proximity and k-core central immune nodes. The effects on the transmission range were observed and the immune performance was compared. It is found that in ER and SF networks, when the adoption threshold is not correlated with degree and is completely negative, it is recommended to choose the degree central immunity, and when the adoption threshold and degree are completely positively correlated, the heterogeneity of the adoption threshold is weak. All of them are suitable for the selection of degree central immunity, but when the heterogeneity of threshold is strong, ER network is suitable for selecting k-core central immune SF network, each central immunity can not inhibit the transmission.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O157.5
【参考文献】
相关期刊论文 前1条
1 王伟;舒盼盼;唐明;高辉;;网络传播动力学模拟方法评述[J];电子科技大学学报;2016年02期
，本文编号：2177872
本文链接：https://www.wllwen.com/kejilunwen/yysx/2177872.html
