Section: Computer Science and Engineering
Tutor: SILVANO CRISTINA
Advisor: GATTI NICOLA Major Research topic
:Online Learning Methods for Price Advertising in E-Commerce ScenariosAbstract:
In the last two decades the use of the Internet as a channel to sponsor and sell products online has been a key growth diver for small and large companies.
In this thesis we focus on a specific advertising scenario, called Price Advertising, whose popularity has grown in the last few years. In this new advertising setting we have a specific product that is promoted by some advertisers who offer it at different prices. The novelty is that the price is shown in the banner: it is a new variable influencing the user's click probability.
The classical advertising scenario has been analysed in details both from a theoretical point of view, by the computational economics community, and from a practical point of view, by big companies in the advertising area like Google and Amazon. Despite the attention drawn from the classical setting, Price Advertising is almost unexplored in the literature. We will study and design automatic techniques to jointly optimize the bid and the price of a good in the Price Advertising scenario. In this context we will use techniques coming from different fields as machine learning, online learning, economics and optimization in order to handle the specific characteristics of the scenario, e.g., the non-stationarity induced by the behavior of users and competitors.