Allocation of Personalized Ads on the Web

• Main Authors: Teng-Kai Fan, 范登凱
• Other Authors: Chia-Hui Chang
• Format: Others
• Language: en_US
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/26587178336944841025

博士 === 國立中央大學 === 資訊工程研究所 === 98 === Web advertising, a form of advertising that uses the World Wide Web to attract customers, has become one of the world’s most important marketing channels. This study addresses the mechanism of Content-based advertising (Contextual advertising), which refers to the assignment of relevant ads to a generic individual web page, e.g. a blog post. As blogs become a platform for expressing personal opinion, they naturally contain various kinds of expressions, including both facts and opinions. Such opinions can be references for decision making of the Web surfers who browses the blog. Thus, if a blog contains negative opinion of some product, it is less likely the ad be clicked. Besides, the web-site owners would be more willing to have ads which are positively related to their contents. Hence, we propose the utilization of sentiment detection to improve Web-based contextual advertising. The proposed SOCA (Sentiment-Oriented Contextual Advertising) framework aims to combine contextual advertising matching with sentiment analysis to select ads that are related to the positive (and neutral) aspects of a blog and rank them according to their relevance. On the other hand, although most contextual advertising scheme considers Web surfers as the ad consumer, bloggers themselves are the constant visitors of their own blogs. Thus, it is worth that the advertising scheme put in the first place the interests or intention of the content creators. Thus, we propose BCCA (Blogger-Centric Contextual Advertising) framework which combines contextual advertising matching with text mining technique to select ads that are related to immediate interests as revealed in a blog and rank them according to their relevance. In addition to general web pages, as social networks are becoming a more interactive platform for social activities, more than 20% of online advertisements are served on social networks. The allocation of advertisements based on both individual information and social relationships is becoming ever more important. In our third advertising framework, we firstly propose the idea of social filtering and compare it with content-based filtering and collaborative filtering for advertisement allocation in a social network. Secondly, we apply content-boosted and social-boosted methods to enhance existing collaborating filtering model. Finally, an effective learning-based framework is proposed for combing filtering models to improve social advertising. We experimentally validate our approach using a set of data that includes real ads, actual blog pages and social networks using metrics of information retrieval. The results indicate that our proposed method could efficiently identify those ads that are relevant to users'' interests.