English  |  正體中文  |  简体中文  |  Items with full text/Total items : 52047/87178 (60%)
Visitors : 8695994      Online Users : 326
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    Please use this identifier to cite or link to this item: http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/111148

    Title: 以分群為基礎之線上拍賣詐騙偵測方法
    Other Titles: An effective fraud detection method based on clustering
    Authors: 詹凱薰;Chung, Kai-Hsun
    Contributors: 淡江大學資訊管理學系碩士班
    張昭憲;Chang, Jou-Shien
    Keywords: 詐騙偵測;分類樹;分群;電子商務;Fraud Detection;Binary Trees;Cluster;e-commerce
    Date: 2016
    Issue Date: 2017-08-24 23:45:12 (UTC+8)
    Abstract: 網路拍賣龐大商機吸引不少投機份子加入,運用詐術獲取不法收益,造成消費者大量時間與金錢損失,嚴重影響電子商務未來發展。面對此問題,學者紛紛提出許多詐騙偵測方法,期能降低消費者損失。然而,面對日新月異的詐騙技巧,這些方法並無法獲得令人滿意的準確率。有鑑於此,本研究發展一套新的動態塑模詐騙偵測方法,期能根據待測帳號的特性,動態建立有效的偵測模型。為此,首先我們將蒐集而得之資料進行篩選過濾,移除具有不合理偏差值之記錄。其後,將訓練資料中詐騙者與正常者進行群聚分析。最後,根據待測帳號與這些群聚的匹配程度,找出最適合之正常者與詐騙者群聚來塑模。為驗證提出方法之有效性,本研究蒐集Yahoo!Taiwan實際交易資料進行實驗。實驗結果顯示,與傳統單一分類樹方法比較,動態塑模確實有助於提升詐騙者或正常者之偵測準確率。此外,本研究提出之方法對於不同類型屬性集,亦具有較穩定偵測結果。
    Online auction attracts a lot of speculators using dishonest tricks to obtain illegal benefits. This causes consumers’ loss, including time and money, and have negative impact on development of e-commerce in the future. For this reason, researchers have proposed a variety of fraud detection method to help users to avoid fraud. However, faced with the evolving fraud techniques, existing methods cannot provide satisfied detection accuracy for consumers. In view of this, this study developments a new dynamic fraud detection method. First, we collect the information from web pages of the Yahoo!Taiwan auction site and filter them with removing outliers. Second, we cluster those data into frauds and non-frauds categories. Finally, finding the best cluster combination of frauds and non-frauds sub-models according to detecting result of test data. To verify the effectiveness of this proposed method, the transaction data in Yahoo! Taiwan are gathered for experiments. In comparison with the single decision tree, the proposed dynamic detection method do help to improve the detection, and have stable detection results for different data set.
    Appears in Collections:[資訊管理學系暨研究所] 學位論文

    Files in This Item:

    File Description SizeFormat

    All items in 機構典藏 are protected by copyright, with all rights reserved.

    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - Feedback