透過您的圖書館登入
IP:3.17.135.54
  • 期刊

使用支援向量機器以及線上問卷調查系統建立自動化信用卡防盜刷系統

Automatic Credit Card Fraud Detection System Using Support Vector Machine and Questionnaire-Responded Transaction Approach

摘要


現代人的消費行為偏重於信用卡的使用,但現今的信用卡盜刷問題常導致持卡者和發卡銀行的困擾。本論文利用mySVM以及線上問卷調查系統作為資料蒐集以及分析輔助工具自行開發自動化防盜刷系統分析使用者的消費行為。以往的研究做法是先做資料蒐集,再將蒐集後的資料加以正規化,接著使用支援向量機器(SVM)作為輔助預測工具。這些傳統的研究流程中,光正規化以及將不同資料格式丟入SVM做測試這些動作消耗掉研究者的大部分時間。因此本系統的作法為初期利用線上問卷調查系統蒐集資料後,建立一套可自動正規化以及自動鏈結SVM做預測分析的人機介面。經由這套自動化系統來檢測消費者的消費記錄,當合理的判斷出有異常的消費記錄時,便可以立即的做處理,以達成防止盜刷的結果。

並列摘要


Credit cards are having more and more of an impact on consumer behavior nowadays. Credit card fraud, in particular, has brought several challenges to both credit card users and the banks. This paper introduces a new system that uses a SVM tool and Questionnaire-Responded Transaction (QRT) to automatically analyze consumer behavior. Traditional research collects data, first normalizes it and finally uses SVM tool to predict the result. In the past, these procedures cost lots of research time. In our system, we first use the questionnaire-responded transaction (QRT) approach to collect data first, and then normalize it and use SVM to automatically predict data result by using SVM tool automatically. Therefore, our system can be used to automatically detect credit card records of users automatically. When credit card fraud occurs, our system can immediately notify the user.

被引用紀錄


張昭威(2010)。運用資料探勘方法建構乳癌預後模式〔碩士論文,朝陽科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0078-0601201112113721

延伸閱讀