近年來，由於網路電話使用成本低廉與便利性而廣受使用者歡迎。然而根據統計，利用此電話平台的詐騙案件亦逐年增加，如何有效偵測或預防此網路電話詐騙的發生，遂益受重視。雖然許多專家提出警示與防範注意的方法，然而各種詐騙手法巧立名目，不斷翻新，令人目不暇及。而實際有效的防範方法模組與應用實務的系統研究卻不多見。因此本研究提出一個利用資料探勘技術來有效篩選異常匿名網路電話日誌紀錄的機制，此系統可以依據網路電話的SIP協定內容來挑選比對已存檔的各類可疑的詐騙模式，只要事先針對已知的詐騙手法進行分類建檔的案例夠多，本研究系統即可過濾出可疑的未經請求知通話來源，防範可能的詐騙案發生。研究成果將可用於提昇通訊系統的安全，並且提供進一步改善整體通話效益應用。

Recently the VoIP calls are becoming increasingly popular due to their advantages in cost and convenience. However, the fraudulent activities which occur within new telecommunications systems are flooding too, how to detect fraud attacking and protect users effectively has become an important livelihood issue. Although the subsequent analysis of fraudulent activities are varied and changes with the emergence of new technologies, seldom are reported and implemented by the forms of commercial activity. Therefore, in this paper, we present a new mechanism with data mining techniques that could be used for effectively identifying anonymous VoIP calls and sifting abnormal behaviors through large amounts of log file data. Based on the SIP protocol content of VoIP calls, all the suspicious patterns are compared with the results of registration server, thus the possible fraudulent activities of unsolicited calls may be extracted. Under suitable manipulation with the study system, it is hopeful to improve the system security and furthermore to support the management performance of telecommunication system.