一套線上拍賣之基因式模糊名聲推理方法

機構典藏 > College of Business and Management > Graduate Institute & Department of Information Management > Thesis > Item 987654321/52152

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/52152

Title:	一套線上拍賣之基因式模糊名聲推理方法
Other Titles:	A genetic fuzzy reputation inference method for on-line auctions
Authors:	張鴻文;Chang, Hung-wen
Contributors:	淡江大學資訊管理學系碩士班張昭憲
Keywords:	名聲管理(reputation management);模糊推理(fuzzy inference);基因演算法(genetic algorithm);線上拍賣(online auction);reputation management;Fuzzy inference;Genetic Algorithm;Online Auction
Date:	2010
Issue Date:	2010-09-23 16:58:00 (UTC+8)
Abstract:	隨著網路的普及化與「宅經濟」的促使之下，網路線上拍賣系統的業績也不斷攀升。以eBay為例，2007年的網拍業務收入為59.7億美元，而在2010年該預估可能將達到88億美元至91億美元，顯示線上拍賣的蓬勃發展。然而，拍賣網站會員人數動輒百萬計，如何協助使用者在下標前挑選合適的賣家便成為重要課題。目前，網拍平台大多使用二元名聲系統(binary reputation system)來管理交易者的名聲，雖然簡單易懂，但很難從中獲得完整的名聲參考資訊。有鑑於此，學者們紛紛提出各種不同的名聲計算方式，以協助使用者挑選合適賣家。面對上述問題，本論文發展了一套線上拍賣基因式模糊名聲推理方法-GFRep。方法中同時考量『商品分類相似性』、『評價時間』、『交易金額』與『評價者可信度』等四種因子，並以模糊推理來計算綜合名聲值。為強化模糊規則的效力，我們利用基因演算法來建立模糊規則庫，對規則的後鍵部與模糊變數的歸屬函數進行最佳化。為驗證系統的有效性，我們蒐集eBay網站上實際的交易資料進行實驗，並與三種不同的方法進行比較。實驗結果顯示GFRep在協助買方挑選合適賣家時，能提供比其餘三種方法更合適的建議。 Because of the universal usage of internet, a name "Otaku Economy" appears. An online auction system''s sales increases continual. Take the leader of this online auction systems-eBay for instance. The revenues of auction sales are 5.97 billion dollars in 2007, and the estimation revenues of eBay will reach from 8.8 billion dollars to 9.1 billion dollars. The above-mentioned facts indicate that those users who use the online auction system to trade with others are increasing. The most important issue are how to choose an appropriate auctioneer and an auction website''s members are tens of thousands that cause the difficulties of management. The auction websites use binary reputation management now to manage the auctioneer''s reputation, but still can''t offer entire reputation information. Because of the drawbacks of binary reputation, many scholars announce different calculation of reputation to help users choosing the right auctioneer. In order to help users solving above-mentioned problems, this thesis develop a suit of Genetic Fuzzy Reputation Inference Method- GFRep for On-Line Auctions .To offer more perfect calculation methods, this research think about three factors including "deal-time" ,"deal-amount" and "the credit of dependability". We calculate total reputation value by fuzzy inference. To strengthen the efficiency of fuzzy rules, We use genetic algorithm to create fuzzy rule base optimizing the rule''s consequent and fuzzy variable''s membership function. To validate the system''s effectiveness, we gather the transaction information from eBay. The result reveals that GFRep is more reliable than other three methods.
Appears in Collections:	[Graduate Institute & Department of Information Management] Thesis

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	314	View/Open

Loading...