利用探勘技術發掘具有異常喊價傾向之競標者

線上拍賣是電子商務經營模式中最受消費者歡迎的交易方式之一，雖然線上拍賣可以帶給參與者更多的購買樂趣，但在交易過程中也存有較多的風險，其中如何避免異常競標者哄抬拍賣商品的價格，是線上拍賣詐欺重要的問題之一。在本篇論文中，我們以競標者之交易資料為探勘的資料來源，每一筆交易資料包含競標者曾經喊價過與購買過的商品項目，我們設計兩個探勘方法，分別從以下兩方面發掘具有異常喊價傾向的競標者：一是設計一個兩階段的探勘方法，從交易資料中找出具有異常喊價傾向的競標者；二是以某一拍賣商品為探勘的目標，設計一個兩階段的探勘方法，找出對此一拍賣商品具有異常喊價傾向的競標者。我們根據所提出的方法，設計與建置一個偵測異常競標者的系統。此探勘結果，對拍賣商品避免被惡意地哄抬競價將可提供非常有用的資訊。

關鍵字

資料探勘；線上拍賣；異常競標者

並列摘要

The online auction is one of the most popular transaction ways in the business models of electronic commerce. Although online auction may provide participants more shopping enjoyment than traditional way, it may introduce potential risk in the transaction process. How to avoid the opportunity of fraud that allows bidders to manipulate the price of merchandise intentionally is an important issue in auction fraud. In this paper, bidders’ transaction data are used as the source of mining, and each transaction data contains a bidder's transaction history and detail information. Two mining methods are presented for discovering the fraudulent bidders. The first task is to develop a mining method that contains two steps to find fraudulent bidders in transaction data. The second task is to treat one item as the target of mining, and to develop a mining method that contains two steps to find the fraudulent bidders for the item. Finally, we design and construct a detecting system to find the fraudulent bidders according to the above methods. The results of the mining can provide effective and useful mechanism to detect and avoid the price of merchandise to be manipulated intentionally.