Bonfring International Journal of Data Mining, Vol. 7, No.
1, February 2017
Money Laundering Identification Using Risk
and Structural Framework Estimation
Dr.G. Krishnapriya
Abstract--- Money laundering refers to activities that
disguise money receive through illegal operations and make
them legitimate. It leaves serious consequence that may lead
to economy corruption. One such problem consisting large
amounts of money transferring through various accounts by
the same person or entity is Money Laundering. Money
laundering scheme is quite a complex procedure. It receipts
some empathetic of the deposit transporting actions at many
phases. Detecting money laundering activities is a challenging
task due we propose a risk model framework in Structural
Money Laundering based on Risk Evolution Detection
Framework (SMLRDF). The connection deceitful deal trails a
sequence of connected money laundering arrangements,
structural money laundering uses sequence matching, social
network investigation, and multifaceted happening processing,
case-based examination. The context that put on case discount
approaches to increasingly lessen the input data set to a
knowingly minor size. The context images the summary data
to discovery couples of communications through common
qualities and performances that are possibly complicated in
ML actions. It then applies a clustering method to detect
potential Money Laundering (ML groups), then the risk model
is used to create a valid and accurate transaction scoring
system to be utilized in an ML prevention system. SMLRDF-
dependent risk modeling, which captures the hidden, and
dynamic, relations among none-transacted entities. SMLRDF
has components to collect data, run them against business
rules and evolution models, run detection algorithms and use
social network analysis to connect potential participants.
Keywords--- Money Laundering, Risk Framework, Risk
Evolution Detection, Structural Estimation.
I. INTRODUCTION
D ATA mining applications are deployed in a wide range of
business fields, especially in financial banking,
telecommunication, and the World Wide Web that have to
deal with the extensive amount of data. Simple database
querying is far from enough for information retrieval in those
business areas. Data mining is used to extract more complex
desired information. The information you want is usually
presented as a pattern. Thus pattern recognition, although not
equivalent to data mining, is generally the framework for data
mining.
Money laundering (ML) is a procedure toward type illegal
profits appearance genuine; this is similarly the process by
Dr.G. Krishnapriya, M.C.A, M.Phil, M.E, Ph.D, Assistant Professor,
Department of Computer Science, Sri Saradha College for Women,
Perambalur.
DOI:10.9756/BIJDM.8313
ISSN 2277 - 5048 | 2017 Bonfring
9
which offenders effort to conceal the true origin and
ownership of the proceeds of their criminal activity.
ML behavioral patterns and ML detection framework
features are essential to ML, but traditional research focuses
on legislative considerations and compliance requirements. All
the method to identify the money laundering to focus on the
neighbor transferring in the pattern. So challenges are often
made to their high false positive rate (FPR) and inefficiency
with large data sets. Detection money laundering is the most
important task for the enforcement directors and finance
ministry also.
Complete money laundering, offenders attempt to adapt
financial profits resulting from illegal doings into a
permissible intermediate such as important speculation or
annuity funds presented in retail or speculation banks. This
type of corruption is receiving more and additional erudite and
appears to consume enthused from the chestnut of medication
trading to backing intimidation and confidently not over
looking individual gain.
Though those rule-based schemes have certain pattern
credit capabilities, they do not have knowledge or
simplification aptitudes and container only competition
designs that they previously know. As novel ML arrangements
industrialized, numerous of these answers were powerless to
expose them, as long as offenders with new streets to avoid
detection and the law. Likewise, the money washing groups
have numerous relatives and business among them. The
problematic is the measurement of suggestion and amount of
traversal happening among source and terminus so that the
foundation of washing might not be recognized.
II. RELATED WORK
A context for evolving an smart, discerning scheme of
anti-money cleaning model to classify money laundering.
Different layers play different roles during the analyzing
procedure [1]. Data of Transaction layer and Account Layer
are submitted from the root bank branches and have composed
the primary sources. Only remote intellect may be resulting
from the viewpoints of together internal crusts [2].
Organization layer and Link layer provide views to take a
comprehensive and aggregate discriminating and analyzing
procedure to all data involved in multiple banks, areas, and
departments, to check, contrast, mine, judge and derive in all
those data collected from separate channels [3]. The following
layers have much more advantages during macro situation
judgment and important cases investigation. Irregularity
discovery uses urbane adaptive replicas to appearance finished
communications, seeing infrequent doings [4].
Bonfring International Journal of Data Mining, Vol. 7, No. 1, February 2017 10

Irregularity discovery replicas analyze purchaser conduct
and deal behavior, unique usual patterns from irregularities
that might designate a high-risk drive [5]. Talented of
observing all types of exchange cleaning doings, an
irregularity discovery organization can save up through the
altering appearance of this class of corruption by determining
new heights of deception and other doubtful doings impartial
as they appear [6]. Happening a combination of gathering and
organization methods for examining ML decorations in an
universal speculation bank.
Consumer conduct in speculation actions is complexs
incenumerous influences effect it, likewise show that by
selecting appropriate scopes, modest DM practices can be
functional composed to detect mistrustful ML belongings in
speculation doings [10]. Hence, in this paper, present a one-
step clustering approach basing on some heuristics from AML
experts to improve the performance of our previous solution in
the term of running time [13].
Somenecessary countermeasures against money laundering
have been proposed, including basic statistical analysis which
constrains the amount of the transactions as well as restricting
their frequency [14]. Other methods that complement these
basic security measures are based on checking every customer
against a blacklist originating from previous investigated cases
and a white list to e.g. avoid mistakes when faced with persons
with the same name. Unfortunately, these and other methods
have proved to be insufficient [15].
III. PROPOSED SYSTEM
The proposed method involves three stages, Which
Includes Preprocessing of Data in Framework, ML Dynamic
Risk Model, and Money Laundering Identification. The
structure generates a network representation of all matching
transactions. Then applies a clustering method to find
suspicious ML communities within the network. It then uses
network-based algorithms to filter out unnecessary accounts
and operations. Finally, the extracted communities are
rearranged, sorted and returned as the output of the
framework.

Developers Interior of data Users

transaction

Structural Money Laundering based on Risk Evolution Detection Framework

Data Pre-
Money Processing ML dynamic Money
Laundering
Transactio Risk Model
n Identification

Data repository

Figure 3.1: Money Laundering Transaction Architecture

A. Preprocessing of Data in Framework Flag (this flag can be either True or False).Truth worthy
The financial transaction and financial accounts both contained white lists are inside the flag when it is a true
having Account-ID (unique numeric ID) and Account Trust condition. The dirty funds are integrated into the financial

ISSN 2277 - 5048 | 2017 Bonfring

Bonfring International Journal of Data Mining, Vol. 7, No. 1, February 2017 11

system through sender account, is called placement.
Intermediate accounts are retransferring the funds for to make
safe the source of funds is called layering. The receiver report
gives funds as clean money called integration. The framework
works as receiving the input data and matching transactions
banks. Finally, we have collected 6000 accounts from
different banks having 12 million transactions. The proposed
method has produced efficientresults and detection accuracy is
also higher.

ML Detection Accuracy
searching. The similarity in deposit and withdrawal amounts
are common attributes for financial (matching) transactions.
Algorithm 120

Detection Accuracy %
Input: Transactional Framework TFW 100
80
Output: Preprocessed TFWL. 60
Step1: for each transaction Tn from TFW 40
Classify separate Qualities framework and add to quality 20
list Ql. 0

QL = 1 (()())
End.
Step2: for each Transactional Dataset TD from TFW
Number of Transactions
If1 ()())then
TFW = TFW .
Graph 1: Shows the Efficiency of Identifying Money
End
Laundering
Step4: stop.
The graph 1 demonstrates the abilityto detect money
B. ML Dynamic Risk Model laundering onsome transaction used. It is clear that the
The dynamic risk model incorporates the static attributes, effectivenessis increased if the size of the transactionis
such as static profiles and dynamic social connection attributes increased. The proposed methodology produces an efficient
of the senders and recipients of the transaction. Consequently, result by increasing the size of the operation.
our risk model consists of a dynamic component and a static
element. MLS are identified, AND from input data streams
relevant attributes are collected and given to the detection ML Detection Efficiency
algorithms.
100
90
Detection Accuracy %

Business Rules: particular MLS is assigned for each MLS

pattern associated with the extracted data. To identify the 80
70
MLS patterns, MLS detectionalgorithmshaving the rules 60
related to specific sector. 50
40
MLS Template: Different subtype combinations in major 30
20
pattern types of templates are used. It will be added to this DB 10
when a new form of MLS is discovered. 0
ML Evolution Model: if the evolution of MLS is within
the accepted trend of our model is Determined.
Algorithm
Input: Dynamic DataDD.
Output: Risk Transitional Analysis RTA.
Step1: read each Data RD
For each transaction from Tn Graph 2: Comparison of Money Laundering Detection
Calculate the Risk, status, Framework, transaction details. Accuracy

RTA= 1 () + . The graph 2 shows the comparison of money laundering
detection accuracy between different methods. It shows
End., clearly that the proposed method has produced more
Step2: Stop. efficiency in money laundering detection.

IV. RESULT AND DISCUSSION V. CONCLUSION

The proposed method has been evaluated using various The proposed Structural Money Laundering based on Risk
transactional set collected from different banking sectors, and Evolution Detection Framework (SMLRD) aims to find
we have separated the accounts which are linkedtovarious potential money-laundering groups among a large number of

ISSN 2277 - 5048 | 2017 Bonfring

Bonfring International Journal of Data Mining, Vol. 7, No. 1, February 2017 12

financial transactions. To improve the efficiency of the

framework, detection accuracy methods such as matching
operation detection and balance score filter are used to narrow
down the list of potential ML accounts. Our risk model factors
in the initial account-opening risk as well as the recent
transactions risks, and it presents a risk score that is valid
within and outside the boundaries of a single financial
institution.

REFERENCES
[1] R. Soltani, U.T. Nguyen, Y. Yang, M. Faghani, A. Yagoub and A. An,
A new algorithm for money laundering detection based on structural
similarity, Conference on Ubiquitous Computing, Electronics & Mobile
Communication IEEE Annual, Pp. 1-7, 2016.
[2] Joined, Fatf-gafi.org-Financial Action Task Force (FATF), Fatf-
gafi.org, 2016.[Online]. Available: http://www.Fatf-gafi.org.
[3] J. Huang, H. Sun, J. Han, H. Deng, Y. Sun and Y. Liu, SHRINK: a
structural clustering algorithm for detecting hierarchical communities in
networks, Proceedings of the 19th ACM international conference on
Information and knowledge management, Pp. 219-228, 2010.
[4] C. Phua, V. Lee, K. Smith and R. Gayler, A comprehensive survey of
data mining-based fraud detection research, arXiv preprint
arXiv:1009.6119, 2010.
[5] C. Alexandre and J. Balsa, A multiagent-based approach to money
laundering detection and prevention, Proceedings of the International
Conference on Agents and Artificial Intelligence, Pp. 230-235, 2015.
[6] U.N.O. on Drugs and Crime, Unodc annual report 2014, accessed,
2015.
[7] E. Androulaki, G.O. Karame, M. Roeschlin, T. Scherer and S. Capkun,
Evaluating user privacy in bitcoin, International Conference on
Financial Cryptography and Data Security, Pp. 34-51, 2013.
[8] I. Miers, C. Garman, M. Green and A.D. Rubin, Zerocoin: Anonymous
distributed e-cash from bitcoin, IEEE Symposium on Security and
Privacy (SP), Pp. 397-411, 2013.
[9] S. Raza and S. Haider, Suspicious activity reporting using dynamic
Bayesian networks, World Conference on Information Technology,
Vol. 3, Pp. 987-991, 2011.
[10] S. Wang, A comprehensive survey of data mining-based accounting-
fraud detection research, International Conference on Intelligent
Computation Technology and Automation, Vol. 1, Pp. 50-53, 2010.
[11] G. Zengan, Application of cluster-based local oulier factor algorithm in
anti-money laundering, International Conference on Management and
Service Science, Pp 1-4, 2009.
[12] A. Antoniou and G. Sinha, Laundering Sexual Deviance: Targeting
Online Pornography through Anti-Money Laundering, Intelligence and
Security Informatics Conference (EISIC), Pp. 91-98, 2012.
[13] D.A. Flores, O. Angelopoulou and R.J. Self, Combining digital forensic
practices and database analysis as an anti-money laundering strategy for
financial institutions, Emerging Intelligent Data and Web Technologies
(EIDWT), Pp. 218-224, 2012.
[14] R. Liu, X.L. Qian, S. Mao and S.Z. Zhu, Research on anti-money
laundering based on core decision tree algorithm, International
Conference on Control and Decision, Pp. 4322-4325, 2011.
[15] N.A.L. Khac, A data mining-based solution for detecting suspicious
money laundering cases in an investment bank, Second International
Conference on Advances in Databases Knowledge and Data
Application, Pp. 235-240, 2010.

ISSN 2277 - 5048 | 2017 Bonfring