基于星形用户社区模型的Twitter广告投放

摘要

Nowadays, the Twitter has huge advertising market, but the individualised advertisements are yet just few. In this paper we put forward an advertising approach based on starlike community model. It uses web crawler to get social network information of Twitter users, and uses factors coefficient algorithm of Gaussian model to process users social network information, and preliminarily sifts out those users being interested in products and having influence both, then builds up the star structure model based on them. In second time selection, the out-degree user core node will be determined and the target starlike sub-graph community will be identified, the out-degree core node of the target community groups is employed as the advertisement carrier to proceed the individualised advertisement. Experimental results show that this advertising approach has higher satisfied degree to community users.%目前,Twitter的广告投放市场巨大,但针对个性化的广告投放却很少,提出一种基于星形社区模型的广告投放方式.采用网页爬虫获取Twitter用户社交信息,利用高斯模型的多因素权系数算法处理用户社交信息,初步筛选出对产品感兴趣和有影响力的用户,并对其建立星形结构模型,二次筛选,确定出度核心节点并识别出目标星形子图社区,将该社区的出度核心节点作为广告投放载体进行个性化的投放.实验结果表明该广告投放方式具有较高的社区用户满意度.