Fuzzy Logic Based Effective Bidding Range Computation and Bidder’s Behavior Estimation in Keyword Auctions Madhu Kumari School of Computer and Systems Sciences Jawharlal Nehru University New Delhi, India e-mail: madhu.jaglan@gmail.com Kamal K. Bharadwaj School of Computer and Systems Sciences Jawharlal Nehru University New Delhi, INDIA e-mail: kbharadwaj@gmail.com Abstract— Keyword auctions are being used to sell the positions along the side of organic results shown by search engine when user types a keyword or a query related to keyword in a search engine . It has been a huge revenue generating arena for search engines since last decade. Irrespective of the great success of these types of auctions there are certain research issues which are still in inchoate state and needs urgent attention of research communities e.g. how much a naive bidder should bid without referring to any complex agents, how much he/she will be minimally charged for the participation etc. In this paper we propose a novel scheme to compute effective bidding range based on fuzzy logic which has threefold advantages. Firstly, it provides bidders with the information of his effective range of bids which can ensure his chances of participation and winning. Secondly, it provides auctioneer with the information about bidders bidding behavior which can help in predicting their revenues and lastly it can enforce the minimum reservation prices in natural way. Experimental results are presented to illustrate working of the proposed scheme. Keywords- keyword auctions,bidding behaviour,reservation prices ,sponsereds search,fuzzy logic. I. INTRODUCTION The Internet economy has been largely affected by the introduction of auction based advertising in the form of sponsored links. Sponsored links are a small number of advertisements (ads, henceforth) that the search engine displays in addition to the standard search results. These ads are arranged in positions top to bottom, typically on the side. Normally, the advertiser pays only when the user clicks on the link (known as pay per click (PPC)). It is a difficult task to set a fixed price for each position because the search queries vary widely and with them the value of the positions. Advertisers in this kind of auctions usually bid (PPC) on various keyword related to their products so it is known as keyword auctions A major task for the search engine is to determine the rules of the position auction, and to select, rank and price the ads that will be displayed to the user, according to that auction. . Hence, typically, auctions are used to determine the prices, and positions of the advertisers. The key Auction Service Provides with most commercial interest, Google, Yahoo! and MSN Live make available to advertisers up to three links above the organic results (these are the mainline slots), up to eight links besides the organic results (sidebar slots) and, more recently, MSN Live even sells links below the organic results (bottom slots). According to a report by Interactive Advertising Bureau and PricewaterhouseCoopers, the keyword advertising revenue reached $8.5 billion in 2007. eMarketer estimates that keyword advertising will consistently account for 40% of the total online advertising revenue for years to come .Apart from the interesting revenue statistics there are number of facts about keyword auctions which makes this type of auctions different from other multi-unit auctions .In pay-per-click auctions, keyword advertising providers often learn about advertisers' abilities to generate clicks, e.g., by observing the click-through rates of the advertisement in the past. Such information has been gradually integrated into the keyword auction designs. For example, in 2003, Google started ranking advertisers by the product of their bid prices and their historical click-through rates. In 2005, Google adopted a more sophisticated ranking scheme that weighs advertisers' unit-price bids by quality scores", which are determined by several factors including advertisers' historical click-through rates, the relevance of the advertisement text, and quality of landing pages. The availability of click-through information also enables alternative minimum bid policies. For example, Google has abandoned their one-size-fits-all minimum-bid policy in favor of a new policy that imposes higher minimum bids for advertisements with low quality score. Brief out line of this paper as follows, we have presented related work in second section In third section we explained the terminologies and keywords related to the proposed scheme. Third section underpins the ides of effective range computation and bidder’s behavior estimation and last section concludes the paper with some future extensions to the proposed scheme. . II. RELATED WORK There numerous research issues pertaining to keyword auction varying from matching of appropriate keyword to 299 978-1-4244-4791-6/10/$25.00 c 2010 IEEE