Adaptive Change of the Web Advertising Campaign Parameters as a Click-Fraud Protection Method

Abstract Click-fraud is the main thread of the online advertising industry. In spite of the fact that search engines constantly improve their traffic-filtering techniques, click-fraud remains a big problem. In this paper authors propose a new method for preventing advertiser's losses and improving both his short and long terms benefits. When click-fraud is revealed advertiser saves money on his account (instantaneous result). However, if it remains unrevealed click-through rate increases and therefore general advertisement rating rises up (future result). The essence of the method is that the parameters of the advertising campaign are changed in dynamic and adaptive way in order to maximize the advertiser's payoff function. Authors utilized game theory concepts for the advertiser's payoff function construction and optimization methods for its extremum searching.