Click Fraud Detection of Online Advertising Using Machine Learning Algorithms

With online advertising quickly growing, click fraud has become a major concern for advertisers. Publishers are paid by the advertisers for each click of their ad on the publisher's website. Click fraud is the act of deliberately clicking on online ads with the aim of generating fraudulent revenue for the website host or to drain the advertiser's budget. Both of these motives are driven by the large amount of money in online advertising, making click fraud an appealing tactic for fraudsters. These fraudulent clicks are generated by scripts that repeatedly click on these advertisements. There are methods to detect whether click fraud is occurring by examining the time in between clicks on a given advertisement. This paper proposes a machine learning-based approach for click fraud detection that can assist advertisers in identifying which clicks are fraudulent. The proposed approach uses various machine learning algorithms, such as logistic regression, random forest, and neural networks. These models are trained on the TalkingData AdTracking Fraud Detection dataset of click logs.