社交网络环境下多信息竞争传播博弈均衡影响最大化的研究
发布时间：2019-01-08 15:19
【摘要】：Web2.0技术的不断完善,助推了移动互联网的发展,各种类型的在线社交网络出现,改变着人们的传统生活方式,人们越来越热衷于使用在线社交网络平台进行信息发布和交流互动,社交网络平台因此成为企业营销推广、信息监管的重要渠道。移动互联网时代,社交网络中影响最大化问题的研究对于网络营销、舆论引导、信息监管等具有重要意义。现实生活中,营销学在商品经济和企业经营哲学中有着举足轻重的作用,如何将商品更大化地推销出去以取得最大化的经济利润对于一个企业甚至个人至关重要。影响最大化问题的提出源于市场营销,在市场营销中,某企业为宣传推销某一新产品,如何有效利用有限的成本选择具有影响力的用户进行产品推广,通过“病毒式营销”(viral marketing)以及“口碑效应”(word-of-mouth)等方式实现产品销售利润最大化的营销推广目标,为重要的研究内容。在信息传播方面,传染病传播SIR等模型为早期经典的传播模型,之后,Richardson和Domingos等人提出了影响最大化问题。在各种影响模型中,被广泛研究的为线性阈值模型(Linear threshold model)和独立级联模型(Independent cascade model),现有的大部分研究是基于这两个模型的扩展,且基于单一信息在影响建模中的算法效率和影响范围研究较多,关于多信息的竞争传播影响机制方面的研究比较少。本文在社交网络多信息竞争传播背景下,进行节点影响力研究,提出了改进的权重占比阈值模型,并将多信息竞争具体化为多产品在社交网络中的影响传播,首次就社交网络中产品营销影响提出了引入成本控制机制的产品竞争博弈平衡影响最大化算法。本文的研究内容主要包括:1.针对带权重社交网络重点研究线性阈值模型。在社交网络信息竞争传播背景下,基于线性阈值模型,提出改进的权重占比竞争线性阈值模型,并对传统度中心性节点影响力评估方法进行改进,提出了一种新的节点有效影响力的评估方法,该方法适用于带权重的社交网络;2.结合实际市场营销情景,引入成本控制机制,模拟以最大化自身利润而非打败对手为目的的两种产品竞争博弈平衡影响情景,提出引入成本控制机制的竞争博弈平衡影响算法;3.基于SIR模型,对节点有效影响力评估方法进行对比分析,验证了基于节点度中心性改进的节点有效影响力评估方法的有效性。通过实验模拟小世界模型,实现竞争博弈均衡过程仿真,验证了问题模型和算法的合理性,为社交网络中产品营销影响传播算法设计提出了新的思路,具有重要的实际意义。
[Abstract]:With the continuous improvement of Web2.0 technology, the development of mobile Internet has been promoted, and various types of online social networks have emerged, which have changed people's traditional way of life. People are more and more keen to use online social network platform to publish and exchange information, so social network platform has become an important channel of marketing promotion and information supervision. In the era of mobile Internet, the study of the influence maximization in social network is of great significance to network marketing, public opinion guidance, information supervision and so on. In real life, marketing plays an important role in the commodity economy and business philosophy. It is very important for an enterprise or even an individual to sell the goods more widely in order to obtain the maximum economic profit. The problem of maximization of impact comes from marketing. In marketing, an enterprise how to effectively use limited cost to select influential users to promote a new product in order to promote a new product. Through "virus marketing" (viral marketing) and "word of mouth effect" (word-of-mouth) and other ways to achieve product sales profit maximization marketing promotion goal, as an important research content. In the aspect of information transmission, the SIR model of infectious disease transmission is a classical transmission model in the early stage. After that, Richardson and Domingos et al put forward the problem of maximizing the influence. Among all kinds of influence models, most of the existing researches on linear threshold model (Linear threshold model) and independent cascade model (Independent cascade model),) are based on the extension of these two models. There are more researches on the algorithm efficiency and influence range based on single information in impact modeling, but there are few researches on the mechanism of multi-information competition propagation. In this paper, under the background of multi-information competition in social networks, the influence of nodes is studied, and an improved weight ratio threshold model is proposed, and the multi-information competition is concretely translated into the influence propagation of multi-products in social networks. For the first time, an algorithm for maximizing the equilibrium effect of product competition game with cost control mechanism is proposed for product marketing in social networks. The main contents of this paper are as follows: 1. This paper focuses on the linear threshold model for weighted social networks. Based on the linear threshold model, an improved weighted proportion competition linear threshold model is proposed under the background of social network information competition, and the traditional method of evaluating the influence of the degree central node is improved. A new method for evaluating the effective influence of nodes is proposed, which is suitable for social networks with weights. 2. Combined with the actual marketing situation, the cost control mechanism is introduced to simulate the two product competition game equilibrium influence scenarios which aim at maximizing their own profits rather than defeating their rivals, and a competitive game balance influence algorithm with cost control mechanism is put forward. 3. Based on the SIR model, the evaluation method of node effective influence is compared and analyzed, and the effectiveness of the method based on node degree centrality improvement is verified. Through the simulation of small-world model, the simulation of competitive game equilibrium process is realized, which verifies the rationality of the problem model and algorithm, and puts forward a new idea for the design of product marketing influence propagation algorithm in social network, which has important practical significance.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP393.09
本文编号：2404791
[Abstract]:With the continuous improvement of Web2.0 technology, the development of mobile Internet has been promoted, and various types of online social networks have emerged, which have changed people's traditional way of life. People are more and more keen to use online social network platform to publish and exchange information, so social network platform has become an important channel of marketing promotion and information supervision. In the era of mobile Internet, the study of the influence maximization in social network is of great significance to network marketing, public opinion guidance, information supervision and so on. In real life, marketing plays an important role in the commodity economy and business philosophy. It is very important for an enterprise or even an individual to sell the goods more widely in order to obtain the maximum economic profit. The problem of maximization of impact comes from marketing. In marketing, an enterprise how to effectively use limited cost to select influential users to promote a new product in order to promote a new product. Through "virus marketing" (viral marketing) and "word of mouth effect" (word-of-mouth) and other ways to achieve product sales profit maximization marketing promotion goal, as an important research content. In the aspect of information transmission, the SIR model of infectious disease transmission is a classical transmission model in the early stage. After that, Richardson and Domingos et al put forward the problem of maximizing the influence. Among all kinds of influence models, most of the existing researches on linear threshold model (Linear threshold model) and independent cascade model (Independent cascade model),) are based on the extension of these two models. There are more researches on the algorithm efficiency and influence range based on single information in impact modeling, but there are few researches on the mechanism of multi-information competition propagation. In this paper, under the background of multi-information competition in social networks, the influence of nodes is studied, and an improved weight ratio threshold model is proposed, and the multi-information competition is concretely translated into the influence propagation of multi-products in social networks. For the first time, an algorithm for maximizing the equilibrium effect of product competition game with cost control mechanism is proposed for product marketing in social networks. The main contents of this paper are as follows: 1. This paper focuses on the linear threshold model for weighted social networks. Based on the linear threshold model, an improved weighted proportion competition linear threshold model is proposed under the background of social network information competition, and the traditional method of evaluating the influence of the degree central node is improved. A new method for evaluating the effective influence of nodes is proposed, which is suitable for social networks with weights. 2. Combined with the actual marketing situation, the cost control mechanism is introduced to simulate the two product competition game equilibrium influence scenarios which aim at maximizing their own profits rather than defeating their rivals, and a competitive game balance influence algorithm with cost control mechanism is put forward. 3. Based on the SIR model, the evaluation method of node effective influence is compared and analyzed, and the effectiveness of the method based on node degree centrality improvement is verified. Through the simulation of small-world model, the simulation of competitive game equilibrium process is realized, which verifies the rationality of the problem model and algorithm, and puts forward a new idea for the design of product marketing influence propagation algorithm in social network, which has important practical significance.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP393.09
【参考文献】
相关期刊论文 前1条
1 刘琳;陈云翔;葛志浩;;基于正态分布区间数的概率测度及多属性决策[J];系统工程与电子技术;2008年04期
相关博士学位论文 前1条
1 刘晓东;大规模社会网络中影响最大化问题高效处理技术研究[D];国防科学技术大学;2013年
，本文编号：2404791
本文链接：https://www.wllwen.com/guanlilunwen/yingxiaoguanlilunwen/2404791.html
