with the popularity of online advertising, pay by per click (pay per click) mode gradually accepted by everyone. But the problem is that the prevention of fraud clicking force in the eyebrow, because it will directly related to the long-term survival of this advertising model and whether it can become a real source of revenue for the site owner.
below describes how the Google Adsense system from the point of view of the system to prevent click fraud, I hope other online advertising system to prevent false clicks can have a good guiding role:
1 click rate = number of clicks / views.
click rate is a key method to judge whether there is no fraud clicks, you can imagine a website advertising click through rate of more than 10% will mean that what.
of Click/ of Viewed
2 click coverage / independent IP, if there is a distribution; a single IP (click / browse) = click coverage exceeded 3 times the system error will be suspected of cheating.
examples, such as from the 18.104.22.168 user browsing 16 page, click 4 ads, but the ad click rate "calculation from 1 to 5%", then calculated:%5 X 16 Sqrt = ~ 1, variance (1) = 1, 4/1 = 4 = click coverage Gauss, according to the distribution of Mathematics, the probability of less than 1/10000.
Ratio IP distribution VS ND
Click on the "click on the coverage rate of
3 / IP /
" time according to the time sequence of the click rate analysis, if there is an obvious peak in a certain period of time, then this will mean that potentially may click fraud.
Ratio VS time O
4 "load time and click ads of the time difference and analysis between every two click time difference sequence, O
load ad Click time and time difference should be a Poisson distribution of possion and distribution, between every two click time difference should also be a Possion distribution, if the time in seconds, more than 25 seconds it basically presents the Gauss distribution shape.
time of loading of distribution VS 4$
time />time click Possion