Identification of Click Fraud and Review of Existing Detection Algorithms

Ad Tech is a growing business, that has already taken over the traditional way of showing ads. It has made the ad-industry better with better reach to audience where the impact of the ads can be measured and acted upon to optimize the performance of the ads. But exposure to internet comes with its own pros and cons. Out of the three ways of paying out money, RPC payout method leads to click-fraud (spam-clicks or clicks with less intent), to increase the count of clicks (eg. Bot clicks) of generate more revenue, eventually affecting the quality of traffic. The paper discusses various ways being used to detect how ad-serving works, ways of detecting spam clicks and monitoring activity of users, and a proposed method. In this proposed method, visitors on the website and their identification details are analysed and each parameter is assigned a weight according to significance. Then each click is assigned a rank which is compared to a threshold value to determine whether the click is suspicious or not. The future scope of this method along with optimisation techniques is also discussed in this paper.