International Journal of Recent Technology and Engineering (IJRTE)

ISSN: 2277-3878, Volume-7 Issue-4S, November 2018

An Exploratory Data Analysis for Loan Prediction

Based on Nature of the Clients
X.Francis Jency, V.P.Sumathi, Janani Shiva Sri

Abstract -In India, the number of people applying for the The primary objective of the bank is to provide their
loans gets increased for various reasons in recent years. The wealth in the safer hands. In recent times, banks approve
bank employees are not able to analyse or predict whether the loan after verifying and validating the documents provided
customer can payback the amount or not (good customer or bad
customer) for the given interest rate. The aim of this paper is to
by the customer. Yet there is no guarantee whether the
find the nature of the client applying for the personal loan. An applicant is deserving or not. This paper classifies the
exploratory data analysis technique is used to deal with this customers based on certain criteria. The classification is
problem. The result of the analysis shows that short term loans done using Exploratory Data Analysis. Exploratory Data
are preferred by majority of the clients and the clients majorly Analysis (EDA) is an approach to analyse the datasets that
apply loans for debt consolidation. The results are shown in summarizes the main characteristics with visual methods.
graphs that helps the bankers to understand the client’s
behaviour.
The purpose of using EDA is to uncover the underlying
Keywords - Loan analysis, exploratory data analysis structure of a relatively larger set of variables using
technique, client’s analysis, financial categories analysis visualizing techniques.

I. INTRODUCTION II. LITERATURE SURVEY

In [1] the researchers analyse the data set using data
The term banking can be defined as receiving and
mining technique. Data mining procedure provides a great
protecting money that is deposited by the individual or the vision in loan prediction systems, since this will promptly
entities. This also includes lending money to the people distinguish the customers who are able to repay the loan
which will be repaid within the given time. Banking sector amount within a period. Algorithms like “J48 algorithm”,
is regulated in most of the countries as it is the important “Bayes net”, Naive Bayes” are used. On applying these
factor in determining the financial stability of the country. algorithms to the datasets, it was shown that “J48
The provision of banking regulation act allows public to algorithm” has high accuracy (correct percent) of
obtain loans.Loans are good sum of money borrowed for a 78.3784% which provides the banker to decide whether
period and expected to be paid back at given interest rate. the loan can be given to the costumer or not.
The purpose of the loan can be anything based on the In paper [2], “loan prediction using Ensemble
customer requirements. Loans are broadly divided as open- technique”, used “Tree model”, “Random forest”, “svm
ended and close-ended loans. Open-ended loans are the model” and combined the above three models as Ensemble
loans for which the client has approval for a specific model. A prototype has been discussed in paper [2] so that
amount. Examples of open-end loans are credit cards and a the banking sectors can agree/reject the loan request from
home equity line of credit (HELOC). Close-ended loans their customers. The main method used is real coded
decreases with each payment. In other words, it is a legal genetic algorithms. The combined algorithms from the
term that cannot be modified by the borrower. Personal ensemble model, loan prediction can be done in an easier
loans, mortgages, auto payments, instalment loan and way. It is found that tree algorithm provides high accuracy
student loans are the most common examples of close-ended of 81.25%.
loans. Secured or collateral loan are those loans that are In paper [3], using R-language, an improved risk
protected by an asset. Houses, Vehicles, Savings prediction clustering algorithm is used to find the bad loan
accounts are the personal properties used to secure the loan. customers since probability of default (PD) is the critical
Unsecured loans are also known as personal or signature step for the customers who comes for a bank loan.
loans. Here the lender believes that the borrower can repay So, a frame work for finding PD in the data set is
the loan based on financial resources possessed by the provided by data mining technique. R- Language has the
borrower. Liquidity risk is the risk that arises from the technique called as KNN (K-nearest neighbour) algorithm
lackof marketability of an investment that cannot be bought and it is used for performing multiple imputation calculation
or sold quickly enough to prevent or minimize a loss. The when there are missing values seen in the data set.
interest rate risk is the risk in which the interest rates priced The paper [4] had used tree model. It helps to find
on loans will be too low to earn the bank money. whether the banking sector people will be able to overcome
the loan problem with their customers. It provides a high
accuracy of 80.87%.
Revised Version Manuscript Received on 25 November, 2018.
Ms.X.Francis Jency, CSE Department, Kumaraguru College of
Technology, Coimbatore, India
Ms.V.P.Sumathi, CSE Department, Kumaraguru College of
Technology, Coimbatore, India
Janani Shiva Sri,C S Department, Kumaraguru College of Technology,
Coimbatore, India

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E1864017519 176 & Sciences Publication
 An Exploratory Data Analysis for Loan Prediction Based on Nature of the Clients

The paper [5] uses decision tree induction algorithm and lakhs. By these criteria, a new column called Category is
found that the algorithm finds a best way to evaluate the derived.
credit risk. To avoid the credit risk, bankers holds the
technique called as “credit score”, where it helps the
lenders to keep note on who are the applicants who will able
to repay the amount or probability of going into the default
risks. The input given for credit evaluation was customer
data, WEKA software, cibil score. The methodology used in
prediction system was problem and data understanding, data
filtering, system modelling and finally system evaluation.
This was done on the banks existing dataset containing 1140
records and 24 attributes. At last the system was tested and
helps the bankers to make a correct decision on whether to
accept or reject the loan approval.
The paper [6] used predictive model technique and
descriptive model technique to predict the loan approval in
banks. In predictive model technique, classification and
regression were used and in descriptive model technique
clustering and association were used. Classifiers also Fig 1. Annual Income vs purpose
implement several algorithms like naive Bayes, kNN
algorithms of R language and regressors implements several Thus, grouping the Category that is high,
algorithms like decision trees, neural networks, etc., To moderate and low.
undergo this prediction analysis, out of all these algorithms,
Inference from the Figure 1 is as follows:
naive Bayes produces a most accurate classifier and the
algorithms like decision tree, neural network, K-NN  People in moderate category seek loan in the higher
algorithms will be more accurate regressors. The main goal numbers.
of the paper is to predict the loan classification based on the  The field debt consolidation shows the highest
type of loan, loan applicant and the assets (property) that distribution.
loan applicant holds. It was found that the decision tree  Low and moderately categorized applicants try for
algorithm gave an improved accuracy of almost 85% on other purpose and car loans equally.
doing the analysis. B. Trust Customer Classification

III. LOAN APPLICANT DATA ANALYSIS

Whenever the bank makes decision to give loan to any
customers then it automatically exposes itself to several
financial risks. It is necessary for the bank to be aware of the
clients applying for the loan. This problem motivates to do
an EDA on the given dataset and thus analysing the nature
of the customer. The dataset that uses EDA undergoes the
process of normalisation, missing value treatment, choosing
essential columns using filtering, deriving new columns,
identifying the target variables and visualising the data in
the graphical format. Python is used for easy and efficient
processing of data. This paper used the pandas library
available in Python to process and extract information from Fig 2. Trust Customers
the given dataset. The processed data is converted into From the Figure 2 it is inferred as follows:
appropriate graphs for better visualisation of the results and
for better understanding. For obtaining the graph Matplot  There are many customers who does not have
library is used. delinquents, has applied for the loan which
A. Annual Income Vs Purpose Of Loan intimates or indirectly conveys that the applicant
In this Figure 1, the X axis represents the purpose of loan has some chances to get approval of the loan as the
i.e. the purpose for which the loan is applied. Debt applicant have no delinquents. The result is about
consolidation, home improvement is some of the purposes. 53.3% of applicants.
High, moderate and low represents the annual income of  And it is also inferred that the number of people
people who fall in the range as below. applied for loan gradually decreases with the
Low represents the annual income of people between the increase in the delinquent months. This shows that,
range of minimum to 10 lakhs and Moderate represents the the applicant has minimum chances to get the
annual income of people between 10 lakhs and 25 lakhs and approval of loans.
High represents the annual income of people above 25

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E1864017519 177 & Sciences Publication
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-7 Issue-4S, November 2018

C. Loan Term Vs Delinquent Months Figure 4 spectacles the repayment period of the loan
versus credit score under various categories by grouping the
derived column credit category and loan term.Figure 4 have
deduced the following:
 Customers with good and very good credit score
prefer for short term payback period in
contradiction to customers with fair credit score.
 People applying loan for first time prefer short term
loans because lender doesn’t do a credit check so
that the applicant avail loans easily.
E. Loan Term Vs Years In Current Job

Fig 3. loan term vs delinquent months

• Figure 3 deals with the customers who can pay the
loan within term period against customers who
cannot repay their monthly depts within the
particular term.
From the Figure 3 it can be concluded that:
• This analysis can find a higher number of
customers who are able to repay without
deliquiates and for short term. Fig 5. Loan term vs Years in current job
• Almost all applicants who are even delinquent
more than 90 months prefers only short term. Figure 5 displays the count of applicants who have
• Applicants who delinquent more than 90 months various years of experience in current job against the term
are less in number and it indirectly conveys that period of repayment of loans.
their loan will never be sanctioned and if its yes, From the Figure 5 it is concluded that:
the applicant will not be able to pay it back
 Applicants who have various years of experience in
D. Loan Term Vs Credit Category the same job claim the loan for a short period of
From the categories poor, fair, good, very good and time.
undefined, the value the credit category is found, as  Also, the applicants who are freshers claim the loan
applicants without credit score fall into the category for a short period of time.
undefined, people have credit score between 300-850  This also infers that, long term loans are borrowed
between these there are some categories such as credit score by people who are yet to start their own business
between 300 and 579 falls in poor, credit score between 580 and can repay only after this business set well and
brings profit.
and 669 falls in category fair and the credit score between
 Long term goals carry a greater risk to the money
670 and 739 is good and above this is considered to be very
lenders and thus not very easily approved by the
good.
banks.
F. Loan Payment Chances Vs Home Ownership
Loan payment chances which have been classified into
canpay, maypay and not payable. From Current credit
balance, subtracting the monthly debt of that person, one can
find whether the person would be able to repay the loan or
not.
By subtracting, if applicants have less than 50,000
balance, the applicant is considered in the not payable
category and applicants having balance between 50,000 and
3 lakhs are considered to be may pay category and above
than that the applicant are considered to be can pay category.
From the Figure 6 it is concluded that,

Fig 4. Loan term vs Credit category

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E1864017519 178 & Sciences Publication
 An Exploratory Data Analysis for Loan Prediction Based on Nature of the Clients

 People living in rent home fall under not payable

category among ownership credentials
 Applicants who have kept home in mortgage apply
loan in highest numbers.

Fig 6. Loan payment chances vs Home ownership

IV. CONCLUSION AND FUTURE WORK

The main purpose of the paper is to classify and analyse
the nature of the loan applicants. From a proper analysis of
data set and constraints of the banking sector, seven
different graphs were generated and visualized. From the
graphs, many conclusions have been made and information
were inferred such as short-term loan was preferred by
majority of the loan applicants and the clients majorly apply
loan for debt consolidation.
This paper work can be extended to higher level in future.
Predictive model for loans that uses machine learning
algorithms, where the results from each graph of the paper
can be taken as individual criteria for the machine learning
algorithm.

REFERENCES
1. A. Goyal and R. Kaur, “A survey on Ensemble Model for Loan
Prediction”, International Journal of Engineering Trends and
Applications (IJETA), vol. 3(1), pp. 32-37, 2016.
2. A. J. Hamid and T. M. Ahmed, “Developing Prediction Model of Loan
Risk in Banks using Data Mining”.
3. G. Shaath, “Credit Risk Analysis and Prediction Modelling of Bank
Loans Using R”.
4. A. Goyal and R. Kaur, “Accuracy Prediction for Loan Risk Using
Machine Learning Models”.
5. M. Sudhakar, and C.V.K. Reddy, “Two Step Credit Risk Assessment
Model for Retail Bank Loan Applications Using Decision Tree Data
Mining Technique”, International Journal of Advanced Research in
Computer Engineering & Technology (IJARCET), vol. 5(3), pp. 705-
718, 2016.
6. Gerritsen, R. (1999). Assessing loan risks: a data mining case study. IT
professional, 1(6), 16-21.
7. Hsieh, N. C., & Hung, L. P. (2010). A data driven ensemble classifier
for credit scoring analysis. Expert systems with Applications, 37(1),
534-545.
8. https://en.wikipedia.org/wiki/Exploratory_data_analysis
9. https://pandas.pydata.org/pandas-docs/stable/
10. https://www.experian.com/blogs/ask-experian/credit-education/score-
basics/what-is-a-good-credit-score/

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E1864017519 179 & Sciences Publication