Log in
|
Help
|
English

English

Engish-Pirate

한국어

Search

Advanced Search

Find results with:

Add another field

Search by date

to

Searching collections:

CCSU Theses and Dissertations

Add or remove collections

Home CCSU Theses & Dissertations Anti-money laundering behavior : reducing the number of non-productive alerts in structuring...

Reference URL

Share

Add tags

Comment

Rate

To link to this object, paste this link in email, IM or document

To embed this object, paste this HTML in website

Anti-money laundering behavior : reducing the number of non-productive alerts in structuring through effective data mining

View Description

Download

small (250x250 max)

medium (500x500 max)

large ( > 500x500)

Full Resolution

Print

you wish to report:

Your comment:

Your Name:

Submit

Cancel

...

Back to top

Powered by CONTENTdm®

Select the collections to add or remove from your search

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z

Select All Collections

C

CCSU Student Publications

CCSU Theses and Dissertations

G

GLBTQ Archives

M

Modern Language Oral Histories

O

O'Neill Archives Oral Histories

P

Polish American Pamphlets

T

Treasures from the Special Collections

V

Veterans History Project

OK

Cancel

Identifier	Thesis 2225
Author	Rivera, Edwin, 1978-
Title	Anti-money laundering behavior : reducing the number of non-productive alerts in structuring through effective data mining
Publisher	Central Connecticut State University
Date	2012
Resource Type	Master's Thesis
Notes	With the rules and regulations the federal government has put in place, financial institutions must consistently improve their anti-money laundering programs. Financial institutions use alert monitoring systems to combat a number of money laundering behaviors. The alert monitoring systems tend to be third-party applications that generate thousands and thousands of weekly and monthly alerts of which a majority do not indicate suspicious behavior. This drives up cost as every alert generated by the monitoring system must be investigated by the research analysts. Structuring, one of the most common money laundering behaviors, generally occurs when illegitimate funds are deposited or moved in a banking system. Because a report must be filled out by the customer for transactions over $10,000, money launderers tend to conduct a transaction or multiple transactions under the threshold amount to avoid creating a paper trail. This study presents a method to reduce the number of non-suspicious alerts for the Structuring money laundering behavior using data mining methods and models with SPSS Clementine/Modeler. The CART and C5.0 decision tree algorithms as well as logistic regression were used to reduce the number of non-suspicious alerts, or false positives. In addition, the voting method of combining models was used to further increase the effectiveness over individual models. Because financial institutions can be heavily penalized in the millions of dollars by missing suspicious activity, the goal is to create a method to remove false positives and true negatives without removing true positives and false negatives prior to the analysts investigating the alerts generated by the alert monitoring system. This research found that no one individual model allowed for this. However, the voting method of combining models found that roughly 5% of the alerts could be removed from being analyzed without losing any suspicious alerts
Subject	Money laundering -- Prevention Money laundering investigation -- Technological innovations Data mining
Department	Department of Mathematical Sciences
Advisor	Larose, Daniel T.
Type	Text
Digital Format	application/pdf
Language	eng
Rating

Find results with:

Searching collections:

Home CCSU Theses & Dissertations Anti-money laundering behavior : reducing the number of non-productive alerts in structuring...

Anti-money laundering behavior : reducing the number of non-productive alerts in structuring through effective data mining

Description

Tags

Comments