20 likes | 167 Views
Senior Project – Computer Science - 2012 Using Agents to Model The eBay Economy David Shepard Advisor – Prof. Barr. Process Objective : Determine if a computer program can create agents to emulate the behavior of real bidders on eBay
E N D
Senior Project – Computer Science - 2012Using Agents to Model The eBay EconomyDavid ShepardAdvisor – Prof. Barr • Process Objective: Determine if a computer program can create agents to emulate the behavior of real bidders on eBay Method: Retrieve detailed data on real eBay auctions, create agents that use this empirical data to model bidders behaviors Success: Simulated auctions and eBay auctions follow a similar pattern • Methods • After failed attempts to collect the necessary data directly from Terrapeak and using the eBay API, resorted to screen scraping • Beautiful Soup module for python • Able to scrape bidding information off of actual bidding summary web pages • Modeling • Retrieved bidding summaries for 200 ipodnano auctions • Filtered bidding by users (approx. 7 bidders per auction, 2 bids per bidder) • Formulated linear and polynomial equations relating a given user’s response to a prior bid • Separated bidders according to two strategies, lowballers and proxy bidders • Simulated auctions with 5-9 bidders (1-3 lowballers, 2-6 proxy bidders) • Each bidder bid according to their respective equation list • Result/Evaluation • Tentative yes, computer agents can model bidders behavior • Correlation is strong, not perfect • Average slope of 100 real-life auctions is within .001 of average slope of 100 simulated auctions • Future work is needed to see if bidders behaviors follow a similar pattern across all eBay auctions