請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/50110
標題: | 可適用於人工智慧數據資料庫之洗錢行為表徵研究 Establishing Characteristics of Money Laundry Behavior for Artificial Intelligence Database Application |
作者: | Tsung-Hsien Wu 吳宗憲 |
指導教授: | 陳達仁(Dar-Zen Chen) |
關鍵字: | 人工智慧,機器學習,資料庫,洗錢,金融情報中心, AI,Machine Learning,Database,Money Laundry,FIR, |
出版年 : | 2020 |
學位: | 碩士 |
摘要: | 國泰民安的一大表現,就是讓人民有安居樂業的環境,而犯罪者犯行的最大動機,就是能順利的使用犯罪所得,否則,犯罪行為將變成毫無意義。國家對於洗錢防制的作為,目的就是阻絕犯罪者順利使用犯罪所得,但是每天全世界有海量的金融交易紀錄,根本不可能以有限的人力去查核出洗錢行為,近年來,人工智慧的應用已經廣泛在各個領域中,雖然人工智慧技術目前尚屬初級,但是在實務上已經有其實用性。
文章內容先介紹洗錢行為的內涵及危險因子的表徵,國際間及我國目前洗錢防制的機構及運作模式,使建置洗錢防制人工智慧的工程師,能理解洗錢防制的邏輯為何,進而可以設計出適用於實務的人工智慧系統。 本文期待,將人工智慧引入洗錢防制中,利用人工智慧對於海量的金融交易紀錄加以查核,並自主不停地擴充、更正資料庫,使洗錢犯行的警示更加精確、廉價。然而,建置足夠效能運用於洗錢防制的人工智慧,在機器學習時,必須由工程師提供足夠且堪用的資料。本文即是將實務中可能表現出洗錢行為特徵的危險因子,區分為名單、地區、可疑金融機構交易等種類,加以列舉出,並希望將此類危險因子建立起的資料庫,供人工智慧利用,期使工程師能建置出最佳的洗錢防制人工智慧系統。 An important mission of a wealthy and secure state is to provide its citizens with a healthy, secure and sustainable living environment. The greatest incentive for criminals to commit crimes is that they can profit from committing a crime. Otherwise, committing a crime would have been meaningless. Hence, an ideal mechanism for a state to prevent money laundering is to eliminate the possibilities for a criminal to benefit from the proceeds of a crime. Nevertheless, there are massive quantity of financial transactions taking places all around the world on a daily basis, and with limited manpower, it is impossible to identify every money laundering activity without assistance of advanced technology. In recent years, the application of artificial intelligence (AI) has been widely use in various fields. Although the AI technology is still in its infancy, it already has its practicability in many applications. This article covers the connotation of money laundering activities, the features of risk factors, the current international and domestic institutions of money laundering prevention, the operation modes to enable the engineers who establish AI for money laundering prevention to understand the logic behind, and finally how to design an AI system applicable to prevention practices. The ultimate goal of this paper is to promote the use of the AI technology in money laundering prevention, i.e. applying AI to examine the massive financial transaction records. The ability for AI technology to continuously, autonomously expand and update the database will enable the detections of money laundering activities to become more accurate, timely and cost-saving. In order to establish an AI technology specifically for money laundering prevention, the engineers must provide adequate and useful data during the course of machine learning. This paper also documented the risk factors while demonstrating the features of money laundering activity in practice, such as listing region and suspicious transactions of financial institution, etc. We hope to establish a database housing such risk factors for the use of AI, with a vision to building the best AI system for money laundering prevention. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/50110 |
DOI: | 10.6342/NTU202002965 |
全文授權: | 有償授權 |
顯示於系所單位: | 工業工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-1108202015395200.pdf 目前未授權公開取用 | 2.74 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。