東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back to Search results for [ null ]

Switch To: Labeled | MARC Mode | ISBD

Linked to FindBook

Google Book

Amazon

博客來

Unsupervised Machine Learning to Create Rule-Based Wire Fraud Detection.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Unsupervised Machine Learning to Create Rule-Based Wire Fraud Detection./
Author:	Barricklow, Austin.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2021,
Description:	33 p.
Notes:	Source: Masters Abstracts International, Volume: 82-12.
Contained By:	Masters Abstracts International82-12.
Subject:	Law enforcement. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28497198
ISBN:	9798505540763

Unsupervised Machine Learning to Create Rule-Based Wire Fraud Detection.
Barricklow, Austin.

Unsupervised Machine Learning to Create Rule-Based Wire Fraud Detection. - Ann Arbor : ProQuest Dissertations & Theses, 2021 - 33 p.

Source: Masters Abstracts International, Volume: 82-12.

Thesis (M.S.)--Utica College, 2021.

This item must not be sold to any third party vendors.

The increase in the number of international wires requires a more efficient way of reviewing activity than the previous reactive approach. Waiting for victims of fraud to report activity before it is investigated often leads to the funds being withdrawn before any type of enforcement action can be taken. The number of transactions occurring each day makes it nearly impossible for even the most well-equipped team of investigators to review all activity. Therefore, there is a need to triage activity and review only the most concerning transactions. This is often accomplished by setting up a rule-based program that flags transactions for further review. If these rules are set up based on anecdotal information, it can introduce bias into the program and diminish the quality of services a financial institution can provide. However, if a rule-based program is built with insight gained from the measurable and repeatable output of a machine learning model, financial institutions can more accurately apply rules and have better success in preventing suspicious activity. This report describes methodologies used to develop an unsupervised machine learning classification model that clusters similar transactions and identifies those that do not fit into clusters, suggesting they are anomalies. The data includes thousands of international wire transactions including a few that stood out from the rest. Wire transaction rules were recommended based on similar characteristics of the anomalies. The first rule holds wires sent to a jurisdiction with a risk rating greater than 6 in an amount greater than $4,000,000.00 for further review.

ISBN: 9798505540763Subjects--Topical Terms:

607408
Law enforcement.
Subjects--Index Terms:

Basel Institute

Unsupervised Machine Learning to Create Rule-Based Wire Fraud Detection.
LDR:02775nmm a2200385 4500 001 2344637
005 20220531064613.5
008 241004s2021 ||||||||||||||||| ||eng d
020 $a 9798505540763
035 $a (MiAaPQ)AAI28497198
035 $a AAI28497198
040 $a MiAaPQ $c MiAaPQ
100 1 $a Barricklow, Austin. $3 3683421
245 1 0 $a Unsupervised Machine Learning to Create Rule-Based Wire Fraud Detection.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2021
300 $a 33 p.
500 $a Source: Masters Abstracts International, Volume: 82-12.
500 $a Advisor: McCarthy, Michael.
502 $a Thesis (M.S.)--Utica College, 2021.
506 $a This item must not be sold to any third party vendors.
520 $a The increase in the number of international wires requires a more efficient way of reviewing activity than the previous reactive approach. Waiting for victims of fraud to report activity before it is investigated often leads to the funds being withdrawn before any type of enforcement action can be taken. The number of transactions occurring each day makes it nearly impossible for even the most well-equipped team of investigators to review all activity. Therefore, there is a need to triage activity and review only the most concerning transactions. This is often accomplished by setting up a rule-based program that flags transactions for further review. If these rules are set up based on anecdotal information, it can introduce bias into the program and diminish the quality of services a financial institution can provide. However, if a rule-based program is built with insight gained from the measurable and repeatable output of a machine learning model, financial institutions can more accurately apply rules and have better success in preventing suspicious activity. This report describes methodologies used to develop an unsupervised machine learning classification model that clusters similar transactions and identifies those that do not fit into clusters, suggesting they are anomalies. The data includes thousands of international wire transactions including a few that stood out from the rest. Wire transaction rules were recommended based on similar characteristics of the anomalies. The first rule holds wires sent to a jurisdiction with a risk rating greater than 6 in an amount greater than $4,000,000.00 for further review.
590 $a School code: 1754.
650 4 $a Law enforcement. $3 607408
650 4 $a Information technology. $3 532993
650 4 $a Artificial intelligence. $3 516317
653 $a Basel Institute
653 $a Correspondent banking
653 $a Data science
653 $a Kprototypes
653 $a Z-score
690 $a 0770
690 $a 0206
690 $a 0489
690 $a 0800
710 2 $a Utica College. $b Data Science. $3 3433220
773 0 $t Masters Abstracts International $g 82-12.
790 $a 1754
791 $a M.S.
792 $a 2021
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28497198