The objective to deliver the right advertisement to the right audience becomes crucial in the contemporary Internet marketing. Various ad servers (advertisement platforms) implement different techniques of sociodemographical and behavioral targeting. This objective is hard to accomplish in the information-poor context of the local ad servers. In this paper we propose behavioral attribute driven smart targeting system. Its structure is based on five main entities – set of attributes, community profiles, Internet users, web sites and web banners - ads. The attributes are pre-defined entities, and they represent the traits of the Internet users, their interests demonstrated through their online behavior. Attributes are chosen from the finite set. The community profiles are characterized by attributes they are assigned with different degrees of membership. The affiliation of a particular Internet user to a community profile is dynamically calculated with the fuzzy logic model, after relatively small number of clicks by the user (to accomplish the dynamic of interests of the Internet users). Web sites are assigned with an attribute, and visits of the user on particular sites determines the community profile of the Internet user. Furthermore, when a new ad is placed, it is assigned to one or more attributes.
Internet marketing, behavioral targeting, ad server; fuzzy logic, fuzzy neural system.