Oleg and Ekaterina will give an introduction to online advertising and real-time bidding, and also will describe typical problems from the industry and how machine learning is used to solve them: prediction of user’s actions ad fraud detection. In the report we will address of types of fraud in RTB ecosystem (bots, ad stacking, spoof sites). Than we’ll discuss what can be caught by means of various algorithms including machine learning, modified bid clustering for good traffic (human) and bad (bot). Also talk about which clustering method is better, way of learning (supervised/unsupervised), feature selection in terms of fighting fraud. As for the technical part we will discuss the area of values of the variables (size of learning sample, number of Google Cloud Engine machines needed) and the possible ways of computational optimization.
