It’s Fraud! Application of Machine Learning Techniques for Detection of Fraudulent Digital Advertising

Due to the on-going COVID-19 pandemic, industries are heavily reliant on online meeting platforms. The pandemic has forced most MNCs to rely on online platforms such as Zoom or Microsoft Teams to hold daily business meetings and even conferences and other corporate events. Schools and other educational institutions have also been forced to conduct classes and all other events through these platforms. This increased unavoidable dependence on online platforms has resulted in an exponential increase in advertising fraud. Advertising fraud is the application of any method or technology that hampers the proper delivery of advertisements to the proper audience or the proper place, or forcefully inserts advertisements at undesirable times or locations. This could take multiple forms and has become far more widespread with increased use of online platforms. Some common methods used for digital fraud can be the pay-per-click (PPC) model, domain spoofing or in the form of bots, but the main objective is to gain financial advantages from advertising transactions. The primary objective of this study is to identify and understand the factors lying behind the presence of fraudulent activities on any online medium and to analyse the probability of downloading an application after coming across the online advertisement, and watching it. The study also aims to highlight how marketing agencies tend to float fraud advertisement just to gain more revenue from their end.