A Smart Targeting System for Online Advertising

Abstract

With the rapid increase of online advertisements in marketing, as well as the user's attention resources are becoming scarce, targeting technique is applied to improve the efficiency and effectiveness of online advertising by delivering the right advertisement to the right audience, at the right time and situation, and with the right method. The attention of current targeting is transformed from content targeting, frequency targeting, time targeting and geographical targeting to the extended targeting based on user’s characteristics, with data mining technique and behavior modeling technique on the analysis of users. The protection of privacy has been an argued issue along with the application of above technique. This paper presents a smart targeting system with high efficiency, effectiveness and protection of privacy. It uses Web content mining and Web usage mining techniques to track and mine the user behaviors hiding in the historical and current user sessions, and designs interfaces for maintaining the advertising rules, which can control the targeting system to automatically deliver the personalized advertisements. All the related knowledge and rules are integrated by one unified vector space model. The process of targeting doesn’t require any sensitive data input by users, so their privacy is protected.