考虑用户信息转发习惯和信息偏好的在线社会网络中信息扩散模型及仿真

Modeling and Simulation for Information Diffusion in online Social Networks via Considering Users’ Forwarding Habits and Preferences

【目的/意义】研究用户转发习惯和信息偏好差异对信息扩散范围的影响,提出相应建议。【方法/过程】提出了一个三类未知用户的信息扩散模型,比较了两类网络中各类节点比例的差异,分析了概率、未知信息用户比例和转发阈值的影响。【结果/结论】度增加时,两类节点的比例在BA网络中先有优势后又变成劣势,最终几乎无差别;未知节点比例有优势时转发概率增加,知道节点比例一直增加,犹豫节点比例先增后减;度较小时,知道节点比例随着喜欢转发信息的未知节点比例减少和转发阈值增加而降低,度较大时结果相反;犹豫节点比例随转发阈值增加而减少。【创新/局限】考虑到信息内容的偏好差异以及转发习惯的差异对信息扩散的影响,但尚未考虑到信息的新颖度随时间变化等因素对转发信息的概率的影响以及转发信息节点对不同未知信息节点的影响力的差异。

【Purpose/significance】The influence of user’s forwarding habits and information preference differences on the scope of information diffusion is studied and puts forward corresponding suggestions in this paper.【Method/process】The new model which includes three kinds of unknown nodes is proposed,compared the difference between WS network and BA network when diffusion process was finished. The effects of probabilities, the proportion of unknown nodes and forwarding threshold are analyzed.【Result/conclusion】As the network degree is increased, the advantage of two types of nodes in BA network is decreased gradually until become disadvantage, and there is almost no difference in the end;only one kind of unknown users occupy the proportion advantage in the network,the forwarding probability of such users is changed have a greater impact on the proportion of known nodes and hesitant nodes;but the proportion of hesitant nodes is increased and then it is decreased with the increase of forwarding probability gradually;when the network degree is smaller, the proportion of known information nodes is decreased as decrease the proportion of unknown nodes who like to forward information and increase the forwarding threshold, but the result is opposite when the degree is bigger. The proportion of hesitant nodes is decreased with the increase of forwarding threshold.【Innovation/limitation】The influence of different preferences of information content and forwarding habits on information diffusion is considered, but not the influence of factors such as the change of information novelty with time on the forwarding probability and the different influence of a spreader node on different unknown information nodes.