手机广告推荐中的用户兴趣建模研究

摘要

为了增强基于WAP网页的手机广告推荐中用户建模的准确性,并对“非邀”式广告推荐中脱离用户兴趣试探性推荐进行修正,针对手机广告推荐中手机屏幕小、用户注意力集中等特点,根据用户对广告的访问历史和操作模式建立其广告兴趣模型和非兴趣模型,同时分析用户网页访问模式探测其网页兴趣度,在此基础上建立用户综合兴趣模型.分别采用基于网页兴趣模型、基于广告兴趣模型和基于用户综合兴趣模型进行广告推荐,随着样本空间增大,综合兴趣模型的查准率明显优于另两者.实验验证了用户综合兴趣模型在手机广告推荐中的有效性和优越性.%For enhancing the accuracy of mobile advertising recommendation on the basis of WAP Web page, and amending the exploratory recommended in the mobile advertising recommendation of which recommendation pattern is No Invitation, on account of mobile phone' s small screen and user' s great attention, builds advertising interest model and noninterest model according the ads visiting history and operation mode. Simultaneously, analyzed the Web access mode to detect the user' s Web interest degree, proposed a user comprehensive interest model which considering the user' s Web interest and user' s ads interest. Employed three recommended patterns to recommend the mobile advertising based on Web page interest model, advertising interest model and comprehensive interest model, with the sample space growing, the accuracy of comprehensive interest model was superior to other two models. Experiments prove that the effectiveness and superiority of comprehensive model in the mobile advertising recommendation.