Modeling Price Dynamics in Online Auctions via Regression Trees

TitleModeling Price Dynamics in Online Auctions via Regression Trees
Publication TypeBook Chapter
Year of Publication2008
AuthorsJank, W., G. Shmueli, and S. Wang
EditorJank, W., and G. Shmueli
Book TitleStatistical Methods in eCommerce Research
PublisherJohn Wiley & Sons
Series TitleStatistics in Practice
ISBN Number978-0470120125

Bidder behavior is central to auction theory. In online auctions this behavior
is largely hidden: while the bid placements of individual bidders is fully observable, bidder arrivals and departures, and bidder strategies are not. When aggregated, bid placements have been shown empirically to possess special properties. Let N(s), 0 <= s <=T, denote the bid arrival process associated with an online auction starting at time 0 and closing at time T. It has been observed in the literature that such a process becomes increasingly intense as the deadline approaches, and often displays a self-similarity property whereby the bid time distributions on the intervals [s, T] become strikingly similar as s approaches T. In this chapter we identify a general process of bidder activity that (under appropriate conditions) generates a bid arrival sequence that possesses one or both of these properties.


Full Text
PDF icon DiffEqTrees-Jank-Shmueli-Wang--rev.pdf402.97 KB

Biblio Tags: