ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL

Title Name: Analysis of various imbalance techniques with data on the real world Research Question: How can machine language aids bank sectors to predict loan default Note: Title name and Research question is amendable according to requirement , please do suggest on Title name and Research Question if needed to change it .

Running head: ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON
THE REAL WORLD
Analysis of various imbalance techniques with data on the real world
Name of the Student
Name of the University

2
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Table of Contents
Abstract............................................................................................................................................4
Chapter 1: Introduction....................................................................................................................5
1.1 Background............................................................................................................................5
1.2 Research Question.................................................................................................................6
1.3 Problem Statement.................................................................................................................6
1.4 Definition of key terms..........................................................................................................6
1.4.1 Importance of Research..................................................................................................6
1.4.2 Motivation of Reviewing Literature...............................................................................7
1.4.3 General findings of Literature Review...........................................................................7
1.4.4 Structure of report...........................................................................................................7
Chapter 2: Literature Review...........................................................................................................9
2.1 Banking Sector.......................................................................................................................9
2.2 Concept of Data Mining......................................................................................................10
2.3 Related Works.....................................................................................................................10
Chapter 3: Research Methodology................................................................................................13
3.1 Research Approach..............................................................................................................13
3.2 Machine Learning Algorithms.............................................................................................13
J48 algorithm.........................................................................................................................13
BayseNet algorithm...............................................................................................................14

3
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
NaiveBayes algorithm...........................................................................................................14
Gantt Chart.....................................................................................................................................14
Sequence diagram..........................................................................................................................16
Activity Diagram...........................................................................................................................17
4. Conclusion and Recommendations............................................................................................17
4.1 Conclusion...........................................................................................................................17
4.2 Recommendations................................................................................................................18
4.3 Future Works.......................................................................................................................18
References......................................................................................................................................19

4
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Abstract
This research has focused on analysis of imbalance techniques with data on the real world. This
research specifies certain issues in banking sector regarding loan prediction. Therefore, certain
issues are identified in the research. These issues can be mitigated by the help of technology
named as Machine Learning. Several machine learning techniques have been identified in the
research. The use of machine learning techniques have been helping in maintaining a keen
approach in the development if the loan prediction system of the bank. Open ended loans can be
obtained for more and more with the help of credit cards. Credit cards are one of the type of
open-ended loans. Customer can have a credit limit that can be bought with both of these types
of loans. Closed Dream Housing Finance Company has been facing challenges in providing loan
to customers. Therefore, Machine learning techniques need to be used in the company. The use
of different categories in the banks have been helping in maintaining keen approach to the
development if the loan management system However, there have been issues faced with
making section to provide loans to individuals. There are three machine learning techniques
described in the research including J48, BayesNet algorithm and NaiveBayes algorithm. These
algorithms have been based on the data mining techniques. Therefore, these techniques might
help in maintaining a keen approach in the growth of the loan prediction system in the company.
There have been proper analysis and comparison done among these three algorithms. This
research will help in identifying the best algorithm that might be used in the Dream Housing
Finance Company in order to automate their loan management system. The use of automate
system will help in enhancing growth of company in the market. There have been some
recommendations provided for enhancing loan prediction model in the company.

5
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Keywords: Machine Learning, Data Mining techniques, neural networks and BayesNet
Algorithm

6
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Chapter 1: Introduction
1.1 Background
Distribution of loans has been core business part of every banks. The main focus of the
bank has been pointed on providing loans to people. The main objective of banking environment
has been investing their assets in safe hands [1]. In recent days, many banks have been approving
loans after a regress procedure of verification and validation of individual. However, there are
still flaws in selecting right candidate for providing loans. This research has focused on how can
machine language aids bank sectors to predict loan default. Banking sector has been facing
issues in providing loans to customers as there have been many fraud cases reported. Banks used
to validate and verify different aspects of customers before allowing loans to them [4]. However,
there are still some errors in the validation and verification processes due to which various fraud
cases have been reported. This research has focused on improving this issue with the help of
Machine Learning. In this case, Dream Housing Finance Company has been dealing with all
home loans. The Company has located in all urban, rural areas and semi urban areas. Customers
used to apply for home loans and after that company used to validate the customer eligibility for
loan. Loan management system has been an important part of the banking system. This help in
increasing the business of the bank in the market [17]. The use of loan prediction system helps in
predicting to whom loan need to be provided based upon the validation and verification.
Machine learning has been playing an important role in automating those prediction in the loan
management system. Various data mining algorithms have been mentioned in the research that
might help in automating loan prediction system in the company. This research has helped in
identifying the best algorithm to be used in the Dream Housing Finance Company to automate
loan management system.

7
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
1.2 Research Question
Following is the research question:
 How can machine language aids bank sectors to predict loan default?
 How can the treatment of the imbalance data using the various imbalance
treatment techniques help in the finance domain to find out the most suspicious
samples and to identify who can be the potential for being the Credit Defaulters
and Loan Defaulters?
 How can these methods impact on the business and how much approximate
amount of loss can be prevented which has to be reported to check the savings?
1.3 Problem Statement
The main problem faced by the banking sector has been providing loans to correct
persons. Loan prediction has been big issue for the banks. Traditionally banks used to select
individual by checking identify proof and other kind of validation. Machine learning is a
technology that might help in predicting loan default. This research will provide a machine
learning model that will help in automating loan management system of Dream Housing Finance
Company [18]. Risks involved in banking sector are related to inability of individuals to
returning back amount of loan and interest to bank. Banks have been facing problems in
selecting customers to whom loan need to be approved. There are various risks in banks
including credit risk in which loan might not be return back to the bank, liquidity risk in which
risks occur due to too many deposits have been withdrawn at instant.

8
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
1.4 Definition of key terms
Machine Learning: Machine learning is an application of artificial intelligence (AI) that helps in
providing ability to automatically learn and enhance rom experience without being explicitly
programmed [6].
1.4.1 Importance of Research
This research has been important for banking sector as this will help in automating loan
management system in the banking sector. This research will help in providing a right way for
the Dream Housing Finance Company in order to automate their loan prediction system in the
market. Utilization of machine learning techniques in the company will help the company on
different ways by automating several systems including loan prediction system. This will also
help in minimizing risks in the banking sector. Therefore, this research has been important for
the Dream Housing Finance Company to maintain their operations in the market.
1.4.2 Motivation of Reviewing Literature
The literature review will focus on analyzing several research papers based on
automating loan department of the banking sector. The literature review will help collecting
several data and information related to the research topic. The literature review will help in
maintaining a keen approach to decide the methodology of the research. Risks involved in
banking sector are related to inability of individuals to returning back amount of loan and interest
to bank. Banks have been facing problems in selecting customers to whom loan need to be
approved.
1.4.3 General findings of Literature Review
The literature review will help in gathering data and knowledge regarding the Machine
Learning techniques for automating loan management system in banking sector. The findings of

9
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
the literature review will help in identifying data ad information related to the impact of data
mining techniques in the banking sector. The use of the data mining techniques in the banking
sector has been helping in maintaining a keen approach in the dependency of the loan prediction
system in the bank system. Data and information related to the loan prediction system in the
bank has been an important issue discussed in the literature review. Therefore, the findings of
literature review will help in identifying the solutions of these issues related to loan prediction
system in the banking sector.
1.4.4 Structure of report
Figure 1: Research Structure
(Source: Created by author)
Chapter 1: Introduction
Chapter 2: Literature Review
Chapter 3: Research Method

10
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Chapter 2: Literature Review
2.1 Banking Sector
There are various risks related to bank loans in type baking sector for those who get loans. The
number of transactions in type banking sector has been rapidly increasing and huge volumes of
data regarding approval of loans [2]. There have been types of loans including Open-ended loans
and Closed-ended loans. Open ended loans can be obtained for more and more with the help of
credit cards. Credit cards are one of type of open-ended loans. Customer can have a credit limit
that can be bought with both of these types of loans. Closed ended loans cannot be loan once that
have been repaid. The balance of the loan get downward when expenditure is made. Secured
loans are loans that always rely on assets. The asset need to be evaluated before having a secured
loan. Unsecured loans might be complicated for having higher concern rates. Unsecured loans
rely on credit history and revenue for meeting criteria for loan [4].
There have been many risks associated with the bank loans, those who get loans and for bank.
Risk denotes the probability of certain outcomes and analyze existing negative threat for trying
to get current monetary operation. Risks involved in banking sector are related to inability of
individuals to returning back amount of loan and interest to bank. Banks have been facing
problems in selecting customers to whom loan need to be approved [19].
2.2 Concept of Data Mining
Data Mining has been an important area of research with aim of extracting data from huge
amount of stored data. Data Mining has been popular in banking sector as there has been call for
efficient analytical methodology for detecting useful information in banks [10]. Skills and

11
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
knowledge have been important necessity for achieving Data Mining task. Data mining process
include three phases:
 Data Preparation
 The actual mining
 Interpretation of the results
Data mining techniques helps in distinguishing between borrowers who pay back loans at
appointed time. It helps in expecting when borrower is at default and whether providing loan to a
particular client will result in bad loans [11]. Customer behavior needs to be analysed related to
being able to repay the loan to the bank. Therefore, data mining helps in analyzing behavior of
customers used in their details and past experiences with banks.
2.3 Related Works
Many researches are conducted that are related to data mining in banking sector. The data mining
applications have been utilized on different data mining techniques including decision trees and
Radial Basis Neural Networks [9]. This study have come up with a way for applying these
applications in the credit risk assessment. Neural Networks features and their fitness have been
granting credit process [7]. According to another research, a complete and systematic way for
producing optimal design of performance model for neural network calculating credit value
based on applications of commercial plan. The neural network has been constructed with their
design that has been able to classify 75 % of loan applicants correctly [8]. The investigation on
neural networks in differentiating loans risks, uncertain logic genetic algorithms, rule simulation
software.

12
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
The adoption of machine learning in the banking sector has been a great approach in automating
loan ad credit management model. As per a research done in the credibility of machine learning
in baking sector, it shows that 80 percent of loan and credit issues have been resolved with the
implementation of data mining techniques in the banking sector [11]. Various aspects of the
sector has been facing issues providing loans to individual. The financial crisis has been faced by
many banks in the due time. Banks have been facing issues with loan distribution to customers as
there have been low probability of returning back loans to the bank [20]. Bans have been facing
risks with credit loans, of-balance sheet risks and technological. However, loan credit risks has
been the most important loss to bank. The use of different categories in the banks have been
helping in maintaining keen approach to the development if the loan management system
However, there have been issues faced with making section to provide loans to individuals [14].
The liquidity in the market has been creating issues as there have been many human errors in the
banking systems. Loan risks has been a potential loss to adverse change in the balance sheet of
the bank. Machine learning algorithms have been helping in intoning keen approach in the
maintaining the security and decisions making system if the banks. The use of machine learning
algorithms has helped in providing security to different functions of the bank. However, it has
focused more in decisions making system of providing loans to individuals. Customers have
been applying for loans to banks and they sued to cross check and validate customer’s data [12].
After that loan is granted to the customer. However, there are issues after providing loan as some
of customers are not able to return back loans to the bank.
However, [16] argued that loan management system of the bank has been appreciated as most of
the loans credit have been positively done without the use of machine learning technology.
Therefore, there might be used of machine learning techniques [21]. The use of machine learning

13
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
techniques have been useful for automating the banking systems in the market. [16] argued on
above statement as machine learning can be used for their courses including automation of
banking systems and generating more positive results with minimal errors. The use of machine
learning techniques in the development of automated banking systems might help in minimizing
errors in the banking system.

14
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Chapter 3: Research Methodology
3.1 Research Approach
Research approach has been of two types including Qualitative and quantitative approaches.
This research will select qualitative approach by conducting experiment with three algorithms.
The issues that are uncovered and the degrees of such issues gets an unfeasible strategy and the
organization of such issues fall under the float of human impediment [20]. Data has been
collected from the experiment and qualitative analysis will be done in the research. The
qualitative approach has heed in maintaining keen approach to the development if the hypothesis
sin the research. A qualitative analysis will be conducted with the experimental data collected
from experimenting with three machine learning algorithms [21]. The three machine learning
algorithms are J48 algorithm, BayseNet algorithm and NaiveBayse algorithm.
3.2 Machine Learning Algorithms
There are various machine learning algorithms. However, this research will focus on three
different algorithms and an analytical comparison will be done to their accuracy in classifying
data correctly in these algorithms.
J48 algorithm
J48 has been an enhanced version of C4.5 algorithms that takes an input set of tables and
generate a decision tree as an output. The output decision tree is like a structure of tree. It also
include root, intermediate and leaf node [16]. The nodes in generated tree include a decision that
guide to outcome. It helps in splitting input data set into mutually exclusive sets and each set
with a label. Splitting measure has applied to determine which attribute lead to the optimal
splitting including information that gain criterion [24].

15
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
BayseNet algorithm
This algorithm depends on the theorem of Bayse [23]. Bayesian network is calculated with
condition probability to all nodes. It helps in directing acyclic graph. Diverse forms of algorithms
have been applied to approximation conditional probability including Genetic Algorithm and K2,
Hill Climbing, Stimulated Annealing, Tabu Search [13].
NaiveBayes algorithm
The Bayesian has been a supervised learning method. It has been characterized with its elegance,
simplicity and robustness. It has become widely used in classification purposes. It supposes that
attributes of a class are self-determining in real life [22].
Technique Correctly classified instance percent
j48 79 %
bayesNet 77 %
NaiveBayes 74 %
Therefore, after classification of data mining techniques algorithms that are j48, bayesNet and
naiveBayes and results, it can be analyzed that the j48 is the best algorithm for loan
classification. J48 algorithm is best as it has high accuracy and low mean absolute error [13]. It
has been capable of classifying instances correctly than other techniques. Confusion matrix of
the three algorithms shows that J48 is the best algorithm [25].
Gantt Chart
Activities 1st
wee
k
1-3
week
s
3-5
week
s
5-7
week
s
7-9
week
s
9-11
week
s
11-13
week
s
13-15
week
s
15-17
week
s
Research

16
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Planning
Developing
Research
questions
Literature
Review
Research
Methodology
Data Collection
Data Analysis
Conclusion and
recommendation
s

17
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Sequence diagram
Figure 1: Sequence diagram of loan predicting model
(Source: Created by Author)

18
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Activity Diagram
Figure 2: Activity Diagram
(Source: Created by Autor)
4. Conclusion and Recommendations
4.1 Conclusion
It can be concluded that with proper analysis of data mining techniques, proper algorithm
can be identified for utilizing in loan prediction system. Loan prediction system has been an
important issues faced in the banks. The data mining applications have been utilized on different
data mining techniques including decision trees and Radial Basis Neural Networks. Therefore,
this research has helped in automating loan prediction system in the Dream Housing Finance

19
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
Company. This company might get enormous profit by installing machine learning techniques in
their processes. This research has focused on how can machine language aids bank sectors to
predict loan default. Banking sector has been facing issues in providing loans to customers as
there have been many fraud cases reported. Banks used to validate and verify different aspects of
customers before allowing loans to them.
4.2 Recommendations
It is recommended that the company need to identify its risks and issues in the banking systems.
The use of Machine Learning techniques in the company need to be implemented. Machine
learning techniques will help in automating loan predicting system in the company. The use of
the J48 algorithm will help in making fast decision in predicting loans. The adoption of machine
learning in the banking sector has been a great approach in automating loan ad credit
management model.
4.3 Future Works
This research has focused on identifying risks and issues in the Dream Housing Finance
Company in providing loans. There are some machine learning algorithms discussed in the
research. A proper comparisons and analysis have been done with these three algorithms. J48
algorithm has been the best algorithm among these three machine learning algorithms. However,
future research might focus on some other machine learning algorithms for analyzing and
comparing them for finding best algorithm. Future researchers might research on more
algorithms that will help in increasing the speed of automating loan prediction system.

20
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
References
[1] Ma, X., Sha, J., Wang, D., Yu, Y., Yang, Q. and Niu, X., 2018. Study on a prediction of
P2P network loan default based on the machine learning LightGBM and XGboost
algorithms according to different high dimensional data cleaning. Electronic Commerce
Research and Applications, 31, pp.24-39.
[2] Goyal, A. and Kaur, R., 2016. A Survey on Ensemble Model for Loan
Prediction. International Journal of Advanced research and Innovative Ideas in
Education (IJARIIE), 2(1), pp.623-628.
[3] Frey, M., Emrich, A., Fettke, P. and Loos, P., 2018, January. Event Entry Time
Prediction in Financial Business Processes Using Machine Learning-A Use Case From
Loan Applications. In Proceedings of the 51st Hawaii International Conference on
System Sciences.
[4] Hardt, M., Price, E. and Srebro, N., 2016. Equality of opportunity in supervised learning.
In Advances in neural information processing systems (pp. 3315-3323).
[5] Chen, J. and Katchova, A., 2019. Agricultural Loan Delinquency Prediction Using
Machine Learning Methods.
[6] Wu, W., 2018. Final Project: Prediction of Loan Default with Machine Learning Method.
[7] Hamid, A.J. and Ahmed, T.M., 2016. Developing prediction model of loan risk in banks
using data mining. Machine Learning and Applications: An International Journal
(MLAIJ) Vol, 3.
[8] Jiang, C., Wang, Z., Wang, R. and Ding, Y., 2018. Loan default prediction by combining
soft information extracted from descriptive text in online peer-to-peer lending. Annals of
Operations Research, 266(1-2), pp.511-529.

21
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
[9] Chen, H., Chen, Z., Xiang, T. and Zhou, Y., 2015. Data Mining on Loan Default
Prediction.
[10] Yu, X., 2017. Machine learning application in online lending risk
prediction. arXiv preprint arXiv:1707.04831.
[11] Rajendran, S. and Srinivas, S., 2017. A Data-Driven Approach for Multiobjective
Loan Portfolio Optimization Using Machine-Learning Algorithms and Mathematical
Programming. In Big Data Analytics Using Multiple Criteria Decision-Making
Models(pp. 191-226). CRC Press.
[12] Mothilal, R.K., Sharma, A. and Tan, C., 2019. Explaining Machine Learning
Classifiers through Diverse Counterfactual Explanations. arXiv preprint
arXiv:1905.07697.
[13] Abdou, H., Abd Allah, W., Mulkeen, J., Ntim, C.G. and Wang, Y., 2017.
Prediction of financial strength ratings using machine learning and conventional
techniques. Abdou, HA, Abdallah, WM, Mulkeen, J., Ntim, CG, & Wang, Y.
(2017)‘Prediction of financial strength ratings using machine learning and conventional
techniques’, Investment Management and Financial Innovation, 14(4), pp.194-211.
[14] Kelleher, J.D., Mac Namee, B. and D'arcy, A., 2015. Fundamentals of machine
learning for predictive data analytics: algorithms, worked examples, and case studies.
MIT Press.
[15] Chen, Y.Q., Zhang, J. and Ng, W.W., 2018, July. Loan Default Prediction Using
Diversified Sensitivity Undersampling. In 2018 International Conference on Machine
Learning and Cybernetics (ICMLC) (Vol. 1, pp. 240-245). IEEE.

22
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
[16] Burrell, J., 2016. How the machine ‘thinks’: Understanding opacity in machine
learning algorithms. Big Data & Society, 3(1), p.2053951715622512.
[17] Witten, I.H., Frank, E., Hall, M.A. and Pal, C.J., 2016. Data Mining: Practical
machine learning tools and techniques. Morgan Kaufmann.
[18] Carmona, P., Climent, F. and Momparler, A., 2019. Predicting failure in the US
banking sector: An extreme gradient boosting approach. International Review of
Economics & Finance, 61, pp.304-323.
[19] Aherne, S., 2018. Using Machine Learning Techniques to Predict a Risk Score for
New Members of a Chit Fund Group.
[20] Ho, S.C., Wong, K.C., Yau, Y.K. and Yip, C.K., 2019. A Machine Learning
Approach for Predicting Bank Customer Behavior in the Banking Industry. In Machine
Learning and Cognitive Science Applications in Cyber Security (pp. 57-83). IGI Global.
[21] Saha, P., Bose, I. and Mahanti, A., 2016. A knowledge based scheme for risk
assessment in loan processing by banks. Decision Support Systems, 84, pp.78-88.
[22] Climent, F., Momparler, A. and Carmona, P., 2019. Anticipating bank distress in
the Eurozone: An Extreme Gradient Boosting approach. Journal of Business Research,
101, pp.885-896.
[23] Boz, Z., Gunnec, D., Birbil, S.I. and Öztürk, M.K., 2018. Reassessment and
monitoring of loan applications with machine learning. Applied Artificial Intelligence,
32(9-10), pp.939-955.
[24] Al-Shawabkeh, A. and Kanungo, R., 2017. Credit risk estimate using internal
explicit knowledge. Investment Management & Financial Innovations, 14(1), p.55.

23
ANALYSIS OF VARIOUS IMBALANCE TECHNIQUES WITH DATA ON THE REAL
WORLD
[25] Leo, M., Sharma, S. and Maddulety, K., 2019. Machine Learning in Banking Risk
Management: A Literature Review. Risks, 7(1), p.29.