Data-driven Ad Resources Planning for Non-guaranteed Targeted Display Advertising
In this project, we propose a publishers to better allocate their consists of two building blocks: inflated Poisson regression model clicking behavior, which directly planning method for ad ad resources. The method (1) an arbitrary-point- to capture users’ ad forecasts the number of clicks, instead of the click-through rate (CTR) as used for event-based auctions; and (2) an optimization model to solve for the optimal ad resource allocation plan, for which we will develop efficient algorithms. The two blocks define each other and are integrated: the output of the forecasting model will serve as the input for the optimization model, and the optimization model actually defines the objective of the forecasting model. In addition, we will study the plan execution problem, i.e., the ad serving procedure. Some extensions to the basic model that allow the planning approach to handle advertisers ‘ specific preferences will also be investigated.
|Effective start/end date||1/05/18 -> …|