• Bidding Behavior in Internet Auction Markets

      Wooders, John C.; Vadovic, Rado; Wooders, John C.; Walker, Mark; Oaxaca, Ronald L. (The University of Arizona., 2006)
      In this dissertation I study bidding behavior in Internet Auction Markets. I focus on practice called "multiple bidding" which occurs when a single bidder places numerous bids throughout the same auction. Multiple bidding appears frequently in the data but the incentives that motivate it are not well understood. In the first chapter I develop a theoretical model in which multiple bidding is an equilibrium behavior by rational bidders. The model has a dynamic auction with two bidders who can search for outside prices while bidding in the auction. Each bidder has a search cost which is her private information. When outside prices are private (independently drawn and identically distributed), then, there is an equilibrium in which bidders with the lower search costs bid only late and always search, while the bidders with higher search costs bid both early and late and search as if they coordinated their search decisions, i.e., the bidder with the lower search cost searches and the other bidder does not. This equilibrium by itself provides an explanation of two frequently occurring bidding patterns (late and multiple bidding). In the second chapter I study experimentally the effect of early bids in dynamic auctions on how bidders search for outside prices. The design has two bidders participating in an ascending clock-auction during which any one of the bidders can pause the auction clock. This I interpret as placing an early bid. Once the auction is paused both bidders can simultaneously search for an alternative outside price. Results indicate that pausing decisions by subjects impact their subsequent searching for outside prices, i.e., whether a subject decides to search or not depends on whether she has paused the auction or not. Subjects behave as if they coordinated their searching decisions: the bidder who pauses the auction also searches with high frequency and the other bidder does not. Because this type of behavior increases both the efficiency and the profitability of the auction we favor the use of policies that promote early bidding in practice, such as, longer auctions and lower public reserve prices.

      Chen, Hsinchun; Zeng, Daniel D.; Huang, Zan; Chen, Hsinchun; Zeng, Daniel D.; Chen, Hsinchun; Zeng, Daniel D.; Nunamaker Jr.; Jay F. (The University of Arizona., 2005)
      Recommender systems automate the process of recommending products and services to customers based on various types of data including customer demographics, product features, and, most importantly, previous interactions between customers and products (e.g., purchasing, rating, and catalog browsing). Despite significant research progress and growing acceptance in real-world applications, two major challenges remain to be addressed to implement effective e-commerce recommendation applications. The first challenge is concerned with making recommendations based on sparse transaction data. The second challenge is the lack of a unified framework to integrate multiple types of input data and recommendation approaches.This dissertation investigates graph-based algorithms to address these two problems. The proposed approach is centered on consumer-product graphs that represent sales transactions as links connecting consumer and product nodes. In order to address the sparsity problem, I investigate the network spreading activation algorithms and a newly proposed link analysis algorithm motivated by ideas from Web graph analysis techniques. Experimental results with several e-commerce datasets indicated that both classes of algorithms outperform a wide range of existing collaborative filtering algorithms, especially under sparse data. Two graph-based models that enhance the simple consumer-product graph were proposed to provide unified recommendation frameworks. The first model, a two-layer graph model, enhances the consumer-product graph by incorporating the consumer/product attribute information as consumer and product similarity links. The second model is based on probabilistic relational models (PRMs) developed in the relational learning literature. It is demonstrated with e-commerce datasets that the proposed frameworks not only conceptually unify many of the existing recommendation approaches but also allow the exploitation of a wider range of data patterns in an integrated manner, leading to improved recommendation performance.In addition to the recommendation algorithm design research, this dissertation also employs the random graph theory to study the topological characteristics of consumer-product graphs and the fundamental mechanisms that generate the sales transaction data. This research represents the early step towards a meta-level analysis framework for validating the fundamental assumptions made by different recommendation algorithms regarding the consumer-product interaction generation process and thus supporting systematic recommendation model/algorithm selection and evaluation.