1: \begin{abstract}

2:   The problem of market clearing is to set a price for an item such

3:   that quantity demanded equals quantity supplied. In this work, we

4:   cast the problem of predicting clearing prices into a learning

5:   framework and use the resulting models to perform revenue

6:   optimization in auctions and markets with contextual information.

7:   The economic intuition behind market clearing allows us to obtain

8:   fine-grained control over the aggressiveness of the resulting

9:   pricing policy, grounded in theory. To evaluate our approach, we fit

10:   a model of clearing prices over a massive dataset of bids in display

11:   ad auctions from a major ad exchange. The learned prices outperform

12:   other modeling techniques in the literature in terms of revenue and

13:   efficiency trade-offs. Because of the convex nature of the clearing

14:   loss function, the convergence rate of our method is as fast as

15:   linear regression.

16: \end{abstract}

17: