Journal of Innovation in Computer Science and Engineering

Mobile advertising gives opportunities for advertisers to only bid and pay for measurable user responses, such as clicks on ads. As a result, click prediction systems are central to most mobile advertising systems. With millions of mobile user's daily activities and active advertisers, predicting clicks on Mobile ads is a challenging machine learning task. This paper presents an experimental study of using different machine learning techniques to predict whether an mobile ad will be clicked or not. Our approach is applied on Avazu mobile Ad click data set. In this paper feature selection is performed to remove features that do not help improve classifier accuracy. Several supervised classification algorithms are applied in experiments and observe that logistic regression with feature selection produces better classification accuracy.