Computational advertising: A data-centric comprehensive web application

Computational Advertising is establishing a new scientific sub-discipline, and brings many disciplines together, including informational retrieval, data mining, statistical analysis, machine learning, distributed systems and so on. It is creating tremendous large market values, and has a mount of real applications, such as online advertising, mobile-based advertising, social network-based advertising and so on. In this paper, we mainly focus on the evolutions of computational advertising, including the evolution of its advertising model (from contract to real time bidding), the evolution of its pricing (from CPM to CPC and to CPT), the evolution of its advertising goal (from branding to direct marketing), and the ecosystem of the computational advertising (e.g., advertiser, publisher, ad network, ad exchange, demand side platform, sale side platform and so on), and the categories of the computational advertising (e.g., banner advertising, sponsored search, ad network and so on), and the algorithms for ad retrieval, ad reordering, and platforms for offline advertising analysis, real time advertising. We find the new trends for computational advertising for the coming years are mobile-based advertising, location-based advertising, and social network-based advertising and so on.