Click Fraud Detection Using Ensemble Classifier

Online advertising fraud is an emerging issue and has drawn researchers’ mind recently because it now poses a significant danger to an online advertising sector. A content publisher signs an agreement to place advertisements on various websites in exchange for commissions on the basis of the amount of traffic these advertisements generate. It may sometimes motivate fraudulent publishers to produce unauthorized clicks on their websites in order to maximize traffic to the advertisements, a practice known as click fraud. Although click fraud is identified as a huge issue, there are not many algorithms that have been able to effectively solve this problem. Our aim in this paper is to overcome the problem of click fraud and come up with an explicit model for the click fraud detection. In this paper, a model is proposed using recursive feature elimination (RFE) and ensemble classifier with Hellinger distance-based decision tree (HDDT). The proposed model is evaluated over the dataset hosted by a Chinese company TalkingData available freely from Kaggle and has resulted in an accuracy of 99.72% as compared to the traditional decision tree.