Prediction of click frauds in mobile advertising

Abstract

Click fraud represents a serious drain on advertising budgets and can seriously harm the viability of the internet advertising market. This paper proposes a novel framework for prediction of click fraud in mobile advertising which consists of feature selection using Recursive Feature Elimination (RFE) and classification through Hellinger Distance Decision Tree (HDDT).RFE is chosen for the feature selection as it has given better results as compared to wrapper approach when evaluated using different classifiers. HDDT is also selected as classifier to deal with class imbalance issue present in the data set. The efficiency of proposed framework is investigated on the data set provided by Buzzcity and compared with J48, Rep Tree, logitboost, and random forest. Results show that accuracy achieved by proposed framework is 64.07 % which is best as compared to existing methods under study.