查詢結果分析
相關文獻
- 以交易記錄的社會網絡結構建立線上拍賣哄抬評價的偵測指標
- Neural Network Procedures for Taguchi's Dynamic Problems
- A Fast and Efficient Competitive Learning Design Algorithm Based on Weight Vector Training in Transform Domain
- 專家系統振動訊號圖型判別之研究
- 反傳遞模糊類神經網路於流量推估之應用
- 類神經網路(Neural Networks)的種類及其在影像處理上的應用
- C++Fuzzy類神經網路物件導向發展系統之建立
- 臺灣汽保費率之估計--對數線性費率模式與類神經網路之比較
- 運用類神經網路於股價指數之套利--以日經225指數為例
- 使用類神經網路預估碳化鎢材料放電加工性能
頁籤選單縮合
題 名 | 以交易記錄的社會網絡結構建立線上拍賣哄抬評價的偵測指標=Detecting Online Auction Fraud of Reputation Inflation through Social Network Structures Embedded in Transaction Records |
---|---|
作 者 | 王俊程; 邱垂鎮; 葛煥元; | 書刊名 | 資訊管理學報 |
卷 期 | 12:4 民94.10 |
頁 次 | 頁143-184 |
分類號 | 496 |
關鍵詞 | 線上拍賣; 社會網絡分析; 信任成本; 哄抬評價; 類神經網路; Online auction; Social network analysis; Inflated reputation; Neural network; Cost of trust; |
語 文 | 中文(Chinese) |
中文摘要 | 線上拍賣漸為一般大眾接受,交易額常創新高之際,卻也慢慢成為網路交易欺騙的主要來源,影響使用者對線上交易的信任。在線上拍賣市場的環境下,由於匿名性的特質,交易者容易取得或更換不同的身份資訊,是以身份難以確認,其個人特性資料不再可以倚賴作為偵測詐欺行為的資訊。由於線上拍賣欺騙的特性在於欺騙者同時具有許多身份,而達到欺騙的目的,必須是這些身份之間有目的之互動。因此在建立線上拍賣欺騙偵測的準則時所採行的策略,必須能夠反應群體互動的特性,如此才能發現其惡意的行為。 過去的研究著重在以不同的方式來認證資訊的價值或真實性,以降低資訊不均,或是平衡因資訊不均所衍生的利益。本研究則以檢視行為者在資訊處理上所耗費的資源為重點,來偵測意圖操弄資訊進行欺詐行為者。運用線上拍賣的大量交易記錄資料為基礎,分析交易者的網絡結構特性,來做為拍賣詐欺行為的偵測指標。研究分析乃由收集美國eBay實際線上拍賣網路資料,以社會網絡分析方法,建構出線上拍賣參與者的交易網絡結構指標--核心結構(k-core)與核心周邊比(core/periphery) ,以偵測呈現出異常交易緊密的群組來分辨可疑的哄抬評價拍賣者。這些指標的辨識能力以二種方式來做驗證。一、統計檢定的分析來證實其顯著性;二、將這些指標與eBay網站提供之交易者評價資訊做比較,分別作為監督式類神經網路的輸入層指標值,來比較兩者所建構之線上拍賣可疑交易者的偵測系統的準確度。兩種偵測系統的比較檢定結果指出,有此社會網絡分析指標者可以提供更高的辨識效果,是以可作為網路拍賣偵測可能詐欺者的先期指標。 |
英文摘要 | Online auctions has become the main frauds source and demolishing the trust to the traders in this well accepted and soaring transaction records internet commerce model. The anonymity of internet transaction in online auction market creates the difficulty for identity verification. Traditional demographic-data-based verifications are no longer valid in Internet auction.. The online auction reputation inflation frauds are utilizing the low cost of multiple identities to build inflated reputation for deceiving. The interactions among these reputation inflated accounts are essential to fabricate the reputation in the online auction market. To detect such premeditated online auction frauds, the effected indicators have to exploit the nature of network community structures of these reputation inflated accounts. To reduce the information inequilibrium or to balance the advantage caused by the information inequilibrium, previous researches have focused on the approach of verifying the genuineness of the information or revealing its value. This research focused on the balancing the information processing capability to unveil the trader who manipulated the reputation. Based on the cumulated magnitude of online auction transaction data, the trader's transaction social network can be characterized as detection indicator of reputation inflated traders. By gathering the actual online auction transaction data from eBay Inc. American and the methodology from social network analysis (SNA), this research are able to construct the transactional network structure measurements to discriminate the reputation inflated traders from regular accounts. The power of these two measurements was verified by two approaches, by statistical significance testing and the accuracy contribution in the supervised neural network model. |
本系統中英文摘要資訊取自各篇刊載內容。