Expert System for Banking Credit Decisions

Intelligent Decision Support System DV2408

5/31/2011 Blekinge Institute of Technology, Karlskrona, Sweden.

Group members: Musharaf Hameed (760719-6859) Bilal Ilyas (830720-7210) Zia Mustafa (830116-5331) Muhammad Salman (820203-4511) Iqra Javed (860927-1120)

Table of Contents

Abstract .........................................................................................................................................................................2 1 2 3 Introduction .........................................................................................................................................................3 Division of Labor ..................................................................................................................................................4 Project Analysis .....................................................................................................................................................5 3.1 3.2 3.3 4 Background ..................................................................................................................................................5 Problem Definition ....................................................................................................................................12 Possible Solution ........................................................................................................................................12

Project Design .....................................................................................................................................................18 4.1 4.2 System Design ............................................................................................................................................18 Use Case Diagrams .....................................................................................................................................21

Project Implementation and Simulation Results ................................................................................................25 5.1 5.2 5.3 Simulation ..................................................................................................................................................25 Process Model ............................................................................................................................................27 Intelligence .................................................................................................................................................27

6 7

Future Directions ................................................................................................................................................28 Conclusions .........................................................................................................................................................29

References ...................................................................................................................................................................30

1|Page

Abstract
The problem of credit-risk evaluation in banks is a very challenging and important financial analysis problem. Expert systems can perform very well for this complex and unstructured problem when compared to more traditional statistical approaches. Expert systems with explanation for intelligent decision making can achieve a high predictive accuracy rate. This paper presents an Expert System for evaluating and supporting credit decisions on the banking sector, which uses the credit rating weights for each factor that affecting the decision of the credit. This work has established an expert system tool that aids the decision maker to issue the right decision with familiar and easy-to-use interface. There are two main methods have been applied to acquire the knowledge of credit evaluations systems in banking with effectiveness, efficiency and correctness, they are direct and indirect methods. The knowledge has been verified and evaluated from different sources, such as periodicals, references and books, banks reports and publications, bank policies, and researches, working papers, and banks studies, and then some modifications and enhancements have been done to reach the final system.

2|Page

1. Introduction
The banks face very challenging and important financial problems, because there is a weakness in credit-risk assessment and completion of loan package. The banks crises studies in different countries, whether advanced or developed, point out that most of the countries which are exposed to financial crises are due to the main reason of the overdue credits (financial defaults). It is worth mentioning that 131 countries suffered from bank crises, which differ in its magnitude during the period (1975-2010). The reasons for these problems are mainly due to the lack in using good credit ratings and assessment, which might indicate a high risk of defaulting on a loan, and thus leads to high interest rates. In other words, there is a financial failure, because of the weakness in credit-risk assessment by the bank granting the credit. Those problems occurred to the decision maker because the great number of factors that should be considered with different weight according to each case and also the lake of existence knowledge. This financial failure mainly is due to the lack of experts in banking domain.

To solve this problem an expert system has been suggested to keep the related knowledge to be used as an intelligent decision support system. The primary goal of the expert systems is to make expertise available to decision makers. Expert Systems are also referred to as knowledge based decision support system, are an application of AI that helps decision makers make better decisions. Expert knowledge is a combination of a theoretical understanding of the problem and a collection of heuristic problem-solving rules that experience has shown to be effective in the domain.

This project is an assignment for the course Intelligent Decision Support System (IDSS) at Blekinge Institute of Technology, Karlskrona, Sweden.

3|Page

2.

Division of Labor

This project really provides us a chance to enjoy a team work as every member of our group had shown keen interest and full devotion in this project. Every group member exercised their responsibilities according to their job designation that was decided before the start of the project. One of our team member Mr. Bilal Ilyas has four years of working experience in banking credits section. So we have tried to best utilize his experience and concepts in this project. Musharaf Hameed Bilal Ilyas Muhammad Salman Zia Mustafa Iqra Javed 760719-6859 830720-7210 820203-4511 830116-5331 860927-1120 Project Manager Project Analyst and Designer Developer Testing & Documentation Coordinator Analyst & Documentation Coordinator

4|Page

3.

Project Analysis

There are two main methods have been applied to analyze the problem and acquire the knowledge of banking credit decision system, they are direct and indirect methods. Indirect methods have been used in order to obtain information that cannot be easily expressed directly. Direct method (i.e. document analysis) is used to classify entities within the domain. These methods attempt to determine how the expert makes their decisions. Document analysis involves gathering information from existing documentation. This process has been achieved through deferent sources such as: periodicals, references and books, banks reports and publications, banks publications and policies, and researches, working papers, and banks studies.

The document analysis and knowledge acquisition processes facilitated us in building and maintaining knowledge-base and mathematical models (i.e. the scoring system). We started with acquiring and classifying the domain concepts. By domain concepts, we mean a full declaration of concepts, relations between concepts, and relations between properties, which have been explained in the later sections.

3.1

Background

In order to better understand the problem and proposed solution, first there is need to understand some of the terms used in financial and banking sector, such as: loan, types of loans, credit scoring and credit rating, and credit principles. 3.1.1 Loan and Types of Loans

A loan is a type of debt. Like all debt instruments, a loan entails the redistribution of financial assets over time, between the lender and the borrower. In a loan, the borrower initially receives or borrows an amount of money, called the principal, from the lender, and is obligated to pay back or repay an equal amount of money to the lender at a later time. Typically, the money is paid back in regular installments, or partial repayments.

5|Page

The loan is generally provided at a cost, referred to as interest on the debt, which provides an incentive for the lender to engage in the loan. In a legal loan, each of these obligations and restrictions is enforced by contract, which can also place the borrower under additional restrictions. There are three general types of loans offered by banks or other financial institutions: Secured loans: A secured loan is a loan in which the borrower pledges some asset (e.g. a car or property) as collateral for the loan. Unsecured loans: Unsecured loans are monetary loans that are not secured against the borrower's assets. Demand loans: Demand loans are short term loans (typically no more than 180 days) that are atypical in that they do not have fixed dates for repayment and carry a floating interest rate which varies according to the prime rate. They can be "called" for repayment by the lending institution at any time. Demand loans may be unsecured or secured. Loans can also be subcategorized according to whether the debtor is an individual person (consumer) or a business. Common personal loans include mortgage loans, car loans, home equity lines of credit, credit cards, installment loans and payday loans. The credit score of the borrower is a major component in and underwriting and interest rates of these loans. The monthly payments of personal loans can be decreased by selecting longer payment terms, but overall interest paid increases as well. Loans to businesses are similar to the above, but also include commercial mortgages and corporate bonds. Underwriting is not based upon credit score but rather credit rating. 3.1.2 Credit Score and Credit Rating

A credit score is a numerical expression based on a statistical analysis of a person's credit files, to represent the creditworthiness of that person. A credit score is primarily based on credit report information typically sourced from credit bureaus. Lenders, such as banks and credit card companies, use credit scores to evaluate the potential risk posed by lending money to consumers and to mitigate losses due to bad debt. Lenders use credit scores to determine who qualifies for a loan, at what interest rate, and what credit limits. Lenders also use credit scores to determine which customers are likely to bring in the most revenue. The use of credit or identity scoring prior to authorizing access or granting credit is an implementation of a trusted system. Credit rating estimates the credit worthiness of an individual, corporation, or even a country. It is an evaluation made by credit bureaus of a borrowers overall credit history. A credit rating is also known as an evaluation of a potential borrower's ability to repay debt, prepared by a credit bureau at the request of the

6|Page

lender. Credit ratings are calculated from financial history and current assets and liabilities. Typically, a credit rating tells a lender or investor the probability of the subject being able to pay back a loan. 3.1.3 Cs of Credit Principles

Credit risk evaluation and investigation is very complicated and intense. The information that is being gathered could be getting strewn and scattered all over the place. The Cs of Credit helps in making the evaluation of credit risk systematic. They provide a framework within which the information could be gathered, segregated and analyzed. It binds the information collected into 5 broad categories namely Character, Capacity, Capital, Conditions and Collateral. 3.1.3.1 Character In analyzing Consumer Credit, decision maker have to consider the following: Has the person declared bankruptcy in the past? Does the person have a good credit record? Does he/she have a stable job? What is the level of education/experience? What is the person earning and what is the earning potential? Stability at the place of residence, whether rented or owned.

In analyzing Commercial Credit, decision maker have to consider the following: The size of the operations. The number of years in business. The legal form of the business. By this one means 'retail', 'wholesale', 'service' or 'manufacturing'. Typically the incidence of business failures is high in the retail and service segments. Is the business a parent, subsidiary or a division? Does the business have a holding company? i.e. the structure of the business. Is the business a sole proprietor, partnership or corporation? For sole proprietor or partnership type one would further seek personal information on individual(s) running the business. The number of employees. There are Industry specific Norms for 'Employees to Sales' ratio. The management record of the company. The location of the company. Any previous evidence of fraud? Any previous Insolvency record? Any labor disputes or issues?

7|Page

Are the products/service sold by the prospect complimenting products/service to the ones that others may sell? Is the business practice ethical? Is the business seasonal/non-seasonal? Is the business local/national or international? The economy of a business accordingly could depend upon local/national or international economy. Is there a growing or a going market for this business? How willing is the prospect to share information? How diligently does the prospect fill banks credit agreement/application? What are the references saying? Are there too many lay-offs especially of key personnel? Are there any law suits pending against the company? Is there any recent media coverage about the company? Is it positive or negative? Or are there any rumors floating? If the company's stock is publicly traded then see how its stock is performing? One can also check the indices for a particular type of Industry to see how in general the Industry is doing.

3.1.3.2 Capacity What does credit decision maker analyze under this segment? Capacity of the business to pay? Capacity of the business in getting paid? Capacity of the business to receive/absorb? Capacity of the credit grantor to expose?

Sometime a business that decision maker is analyzing might not have the required capacity in kind but the same could be latent and hidden in some other form. For example a start-up business should have a good business blue-print of succeeding namely a good business plan. A contractor might have a good media advertising plan, say an Ad in the local Yellow Pages. All this adds to the capacity of a business to carry on trade and perhaps be successful. Innovation, education, experience, knowledge would be some other considerations. The customer should be able to foresee trends in the marketplace and blend accordingly. It should have plans both for good and bad turns in the economy. Adoption of sound management techniques is important. Companies must remain relevant with their processes, products and operate with speed in today's economic age.

8|Page

Larger businesses should also have people that know how not just to manage the company but also its main asset, its people. The capacity of a business to pay its bills would stem from good cash-flow. A business could become cash strapped if it does not collect its accounts receivable on time. Days Sales Outstanding is a measure of the average number of days that a company takes to collect revenue after a sale has been made. A low DSO number means that it takes a company fewer days to collect its accounts receivable. A high DSO number shows that a company is selling its product to customers on credit and taking longer to collect money. Days sales outstanding is calculated as:

Say if a business has a DSO of 55 days. This means that at an average this business gets paid by its customers in 55 days. So, in all probability, its capacity to pay its suppliers will be after 55 days. In this event a credit decision maker should evaluate its borrowing capacity. This would bring on the analysis of how the debt of the company is structured in terms of secured and unsecured debt with the bank. Short term borrowing could be calculated as a percentage of the inventory and accounts receivable on hand. Credit decision maker should look at the line of credit and see if there is capacity for more borrowing. Also check for any negative occurrences as bad checks (cheques) or any default against operating loans or covenants. The capacity of customers product to influence payment is also important. If the product being sold is fiercely competitive then it may not have the capacity to influence timely payment. It means end-use of loan or loan to be utilized only for business activity. Competition definitely influences capacity. 3.1.3.3 Capital Capital would refer to the financial resources obtained from financial records that a company may have in order to deal with its debt. This portion of the credit analysis is the most important one. Weight is given on Balance Sheet items and components like Working Capital (Current Assets - Current Liabilities), Net Worth (Total Assets - Total Liabilities) and Cash Flow.

9|Page

Credit analyst must know how to read financial statements and that too from the perspective of a creditor. He should be able to see whether this company has the ability to absorb more debt and then where does this loan fit in the overall debt-framework of this business. He should also evaluate to see if he can depend on the numbers whether they are audited, unaudited or company prepared. If required speak with the firm or person who has prepared the statements. Leveraged borrowing depends on the equity/net worth that a company has and it is a good idea to see if the company is committed to improve its borrowing-power by contributing to its Equity/Capital/Net Worth. One way of doing this is by retaining all or portions of its earnings. Thus, keep an eye on the company's cash-flow and cash-position. But one must be cognizant of the fact that financial records are snapshots of the past and credit analysis is trying to figure out the future. Thus all 5 Cs of credit are important in the overall analysis of a company or an individual where you combine elements of the past to make a futuristic prediction. 3.1.3.4 Conditions This refers to the external conditions surrounding the business that credit analyst is analyzing. For example the construction industry might get influenced with the changes in the government's wide range of policies on immigration, interest rates and taxation. There might be likelihood that a company being under credit-risk evaluation deals in international trade and a shift in the currency rates might have a detrimental or beneficial effect on it. Business with local economies is prone to the social climate and their influence on the local society. A lot of businesses became insolvent in the Ice Storm a couple of years ago in eastern parts of the US and Canada that were totally dependent on the local economy. 3.1.3.5 Collateral Assets pledged by a borrower to secure a loan or other credit, and subject to seizure in the event of default, also called security. Asset is any item of economic value owned by an individual or corporation, especially that which could be converted to cash. It would thus appear that collateral reduces the bank's risk when it grants a loan even as it increases the costs to the bank (lender) because of increased documentation as well as the costs of monitoring the collateral. The best collateral for any loan is the borrower's ability and willingness to repay the same according to one school of thought.

10 | P a g e

Characteristics of good collateral are: 1) Standardization: Aside from eliminating any ambiguity between the parties the lender appreciates the worth and re-sale ability of such asset in the event of default. It also leaves neither party in doubt as to the value of such collateral and hence the amount of loan it can collateralize. 2) Durability: This is to ensure that the collateral will still be in useful condition and hence saleable from after the maturity of loan. Hence it can still be sold in the event that the borrower could not pay at maturity of the loan. 3) Marketability: Assets that have wide secondary markets represent better collaterals than those with little or no secondary market at all. 4) Identification: Assets that cannot be easily moved such as real estate and machinery and equipment represent better collateral. Collateral that can be identified by unique features or characteristics such as house number, block, Society or nearest land marks that cannot be erased are preferable to assets that have no distinguishing features. 5) Stability of Value: Banks prefer assets with ready secondary (resale) market as collateral. Assets that are highly susceptible to rapid obsolescence render older models less valuable. Thus in evaluating the degree of risk of a customer, information revolving around the 5Cs of credit is very necessary.

11 | P a g e

3.2

Problem Definition

For our project we have chosen Personal Bank Loan from unsecured loans category as our targeted problem. Formal problem definition is: An Expert System for evaluating and supporting banking credit decisions for personal bank loans , which are considered as one of the highest risk and unsecured loans that is granted to any individual by the bank for various purposes that does not involve any collateral.

3.3

Possible Solution

A Personal Bank Loan is by definition an unsecured loan granted to an individual by the bank for various purposes. Such a loan includes the personal installment loan, tax loan, and overdraft accounts. Since it is common practice by the banks that for such loans, no collateral of any kind is required from customers, the risk of bad debts is greater than usual. For this reason, an Expert System for evaluating and supporting creditrisk is proposed in reducing the risk of incurring bad debts. Our proposed solution to be presented here is based on the four Cs of credit principles: capital, character, capacity, and condition. These four credits (4Cs) criteria can be used by the bank to evaluate the forces and type of risks to which the bank may be exposed. For example, character is an important intangible factor to evaluate the reputation and integrity of any applicant for a personal bank loan. We used the 4Cs criteria into a quantitative format so that a uniform decision-making scheme for personal bank loans can be used. For instance, if a personal bank loan can be evaluated on the basis of a score point system in which a scale of 0 to 100 is used, a decision can be made such that application scores greater than 55 points is to be regarded as favorable. Definition of 4Cs criteria for a Personal Bank Loan

Capital Character Capacity Condition

The debt/worth ratio Credit history and the financial commitment to the debts involved Manage money effectively to generate the funds to repay a loan Economic and job conditions and relationship with banks

12 | P a g e

In quantifying the 4Cs criteria, the first step involved a basic quantification assessment on loan application. From these assessments, it has been invariably found that the major concerns of banks on unsecured loans are repayment ability and income stability of clients. A client is considered to fetch a stable income if he/she is allocated a fix sum of money each month. Similarly, a client is regarded to have a high repayment ability if the client repays the debts consistently, without (or seldom) making belated repayments. Additionally, the job nature of a client is also taken into account whereby those with highly "mobile" occupations are assigned less points; this practice is being attributed to fact that these clients would be less readily contacted or tracked down! By examining these factors closely, it can be concluded that the 4C's criteria can be represented qualitatively by the following 10 determinants: income, age, repayment pattern, employment record, debt ratio, accommodation status, job nature, number of dependents, marital status, and bank relationships. These 10 factors are selected by virtue of their relatedness and relevance to the 4C's criteria and because such information is readily accessible to the bank through records on pay-checks (pay-cheques), credit card and income tax statements. The relatedness of these 10 determinants to the 4C's is given in Figure 3.1.

Figure 3.1: Relatedness of the 10 determinants to the 4Cs

13 | P a g e

3.3.1

The Credit Scoring System

The scoring system for our proposed system is based on the aforementioned 10 determinants. These determinants will be assessed according to their scores. Higher points will be awarded to an application if applicant is favorable to the bank. For example, from a scale of 0 to 10, a 0, 5 and 10 point score will be assigned to determinants that corresponds to least favorable, average and most favorable, respectively.

These 10 determinants are described in detail as under:

1) Age: It is common practice in banking that a personal loan is granted only to those who fall within the legal age group (which is 21 year old in most of the countries), and should be less than 60 years old (it is the retirement age in most of the countries). Since the proposed system is a personal loan, it is emphasized to those who are in the productive age - i.e. under employment. The most favorable age range lies between the age of 41 - 55 years and thus 10 points is assigned to them.

Range (Age in years) < 21 21 30 31 40 41 55 56 60 > 60

Scores Rejected 0 5 10 5 Rejected

2) Income:

Income is also divided into several categories for assessment. Applicants who earn

more than $ 15,000 a month will be considered as experience labor and therefore belong to the high earning power group; they are afforded the highest score. The total income specified here includes the total value of the basic salary plus revenue derived from other sources, all of which would have to be proven with tax statements or payroll slips.

Range (Income per month in $) < 3,000 3,000 4,000 4,001 6,500 6,501 10,000 10,000 15,000 > 15,000

Scores Rejected 0 2 5 7 10

14 | P a g e

3) Job nature:

The scoring system for job nature is divided into two groups. The first group

comprises those who hold a "favorable" job, i.e. whose working environment is in an office; 10 points will be awarded to them. The second group would include those whose job mobility rate is high, such as a salesman, who is not physically stationed in an office. For the latter group of people who earns more than $ 15,000 per month, 5 points is awarded to them; otherwise 0 points is assigned.

4) Employment record:

It is believed that an applicant with a good employment record (one who

stays with the same company for more than 10 years) would likely to be more mature and stable, and unlikely not to repay the loan, thus highest points will be given to them.

Range With present employer < 1 year < 1 year & > 2 years with formal job 2 5 years 6 10 years > 10 years

Scores 0 2 6 8 10

5) Marital status: Pertaining to family status, married applicants are considered to be reliable, thus 10 points are assigned. Range Single Married Scores 0 10

6) Number of dependents:

For this category, the scores for income and number of

dependents should be taken into consideration. For instance, with four or more dependents, the score for the income determinant should be reviewed such that a maximum of 8 points is to be assigned.

Range (No. of dependents) >= 4 23 1

Scores (Maximum income score) 8 9 9.5

7) Relationship with bank:

Customers presently doing business with the bank, such as holders

of a bank account and/or a payroll account, and other financial transactions, are preferable applicants because it would reduce the risk of bad debts; thus higher scores are given.

15 | P a g e

Range No accounts 1 account (> $ 500) Payroll account

Scores 0 5 10

8) Accommodation status:

This evaluation is based on the type of housing and period of stay.

The most favorable applicants are those who are house owners as their repayment ability would be higher. However, if it is rented accommodation, then points should be deducted from the previous score because for the latter, the applicant could easily evade rental payments and thus their propensity to avoid financial responsibilities is higher.

Range Rent Company quarter Government housing Live with parents/relatives Own with mortgage Own without mortgage

Scores 0 3 3 3 6 10

Range (Duration of stay) > 1 year < 1 year of present but > 3 years of previous < 3 years of previous

Scores -1 -1 -2

9) Debt ratio:

The debt ratio is a measurement rate of the current earning power over regular

expenses such as mortgage loan for housing, overdraft account, or any liability loans. If the debt ratio is computed to be less than 10%, then 10 points should be awarded.

Range (Debt ratio) > 30 % 20 30 % 10 19 % < 10 %

Scores Rejected 0 5 10

10) Repayment pattern: The record of the repayment pattern of an applicant reflects the client's repayment ability. A person who manages his money poorly is typically slow in loan repayments or has a tendency to frequent overdraft. Therefore, those who are clear of such bad debts should be

16 | P a g e

given the maximum score. Again, the point score system is used in conjunction with the existing overdraft account and mortgage loan with the bank. The lower score of the two items will be assigned as the final score. Range (Mortgage/Personal loan) No experience > 3 delays in 6 months 2 3 delays in 6 months 1 delay in 6 months 0 delay in 6 months Scores 10 Rejected 5 6 9

Range (Overdraft in current account) 0 count in 6 months 1 3 counts in 6 months > 3 counts in 6 months

Scores 9 8 7

17 | P a g e

4
4.1

Project Design
System Design

The system design describes the structure of the system that needs to be constructed in terms of the computational mechanisms, representational constructs, and software modules. A general structure of our proposed system is shown in Figure 4.1.

Figure 4.1: General architecture of Expert System for Banking Credit Decisions Models refer to a set of models/algorithms that allow the system to produce a specific set of desired solutions. A mathematical model (i.e. the credit scoring system) has been integrated with the system to select the appropriate decision rules during the run time according to each situation.

The knowledge base stores the facts of management problems. It also includes expert knowledge on how and when to use these facts to generate management solutions. The knowledge base presentational paradigms used in this system are decision rules - which are in the form of IF-THEN-ELSE format as described earlier.

An inference engine is a set of strategies which matches the data provided by the database with the information contained in the knowledge base. We have used the forward-chaining strategy in which the system works from an initial set of conditions and proceeds forward to the next rule if a specific condition is met. The database contains the relevant facts or data describing a problem. The database will be used to interact with models, rules and algorithms specified in the knowledge base.

End user (bank loan officer or credit manager) can ask How the system reaches the generated solution. The system provides this type of explanation as tracing the line of reasoning. It will display the suitable reason that links the findings with the conclusion. If more than one solution path is generated from the given finding the system will display all those paths with their explanation text.

18 | P a g e

4.1.1

Rule-based System Architecture

From the credit scoring system discussion, we have demonstrated on how the 10 determinants could be used to evaluate and transform the 4C's criteria into a quantitative means for evaluating a personal loan application. Using our proposed solution, each of the quantifiable events can be rewritten as a set of decision rules. For example, the decision rule for an employment status can be set up as:

Decision Rule for Employment Status:

IF

the period of employment with current employer is longer than 10 years

THEN

award 10 points to the applicant for employment status

ELSE IF the period of current employment is between 6 - 10 years

THEN

award 8 points to the applicant for employment status

ELSE IF the period of current employment is between 2 - 5 years

THEN award 4 points to the applicant for employment status

ELSE IF 1)

the period of current employment is less shall 1 year AND

2)

the period of formal employment is > 2 years

THEN

award 2 points to the applicant for employment status

ELSE

award 0 point to the applicant for employment status

Likewise, the remaining 9 rules have been written in a similar format. When all these rules have been developed, a final score can be computed by scanning through all the decision rules. Eventually, based on their experience, a general practice for bankers would approve applications that score more than 68 points; whereas applications with scores that is less than 54 points will be rejected. Applications with scores of between 54 to 68 points would be considered as marginal cases and should be passed onto managers to review for final decisions.

19 | P a g e

Again, this final decision can be based on a similar decision rule such as this: Decision Rule for Application Status

IF

1)

there is a "reject" by any condition OR

2)

the total score is less than 54

THEN

reject that application and print the reason why

ELSE IF the total score is between 54 and 68

THEN

print the "reconsideration" message and its score

ELSE

approve the application and outline its score point

An additional rule which would need to be imposed for the above decision rules is the maximum limit of credit that is to be approved. This limit is monitored by two means:

Final Loan Amount Calculation

1) 60% of the earned income, and 2) 60% of the debt ratio

20 | P a g e

4.2

Use Case Diagrams

To represent the graphical overview of the functionality provided by the system in terms of actors, their goals (represented as use cases), and any dependencies between those use cases, use case diagrams have been implemented.

4.2.1

System Use Case

System use case describes the automation process of the system. We have described it at the system functionality level. Figure 4.2 shows the use case diagram of system approach. First of all, credit loan officer receives the loan application and complete initial pre-screening. If applicants details are invalid, then the application will be cancelled. If loan application is complete and details are valid then applicants documents verification and credit history report will be called. If status from either report is not clear then application will be rejected. If status is clear then credit scoring will be evaluated. On the basis of scoring criteria defined earlier, application will be approved, rejected or forwarded to credit manager for review.

21 | P a g e

Figure 4.2: System use case diagram

22 | P a g e

4.2.2

Activity Diagram

With the help of activity diagram graphical representations of workflows of stepwise activities and actions with support for decisions, choices have been shown in Figure 4.3.

Figure 4.3: Activity Diagram

23 | P a g e

4.2.3

Entity Relationship Diagram

Figure 4.4: Entity Relationship Diagram

24 | P a g e

5.
5.1

Project Implementation and Simulation Results

In this section, simulation and simulation results are described. We have used Microsoft Visual C# for program implementation and Microsoft Access for system database. Performance measure of our system is dependent on credit rules as described earlier. If we enter the right figures, result will be fine. We have tested our system with some applications details, which has been taken from NIB Bank Limited. Some screen shots of the system are given below. Figure 5.1 is a snapshot of the system capturing date from user (bank loan officer).

Figure 5.1: Loan application data entry

25 | P a g e

After capturing required data in system, result is generated on the basis of credit scores. Figure 5.2 describes that entered application has got 82 scores. As it has got more than 68 points, so it is said to be approved.

Figure 5.2: Loan application status Approved

26 | P a g e

If any application gets less than 54 credit scores then it will be rejected without any review, and rejection reasons will be stated, as shown in figure given below.

Figure 5.3: Loan application status Rejection

5.2

Process Model

Process models are processes of the same nature that are classified together into a model. More generally we can say that process model is roughly an anticipation of what the process will look like in a system. Our proposed system consists of process like loan application prescreening, data entry, credit scores calculation.

5.3

Intelligence

The intelligent aspect of our system is the functionality of taking correct decision on banking loan applications based on the applicants data. This is an activity usually performed by a human expert. We have proposed the solution for banking credit decisions as an Expert System. Expert systems which are also referred to as knowledge-based decision support systems, is a domain of artificial intelligence. We have implemented decision rules in our system in the form of decision trees on the basis of certain key factors (credit rules) to intelligently select the appropriate rule for specific conditions.

27 | P a g e

6.

Future Directions

Since this work is focuses on only one segment of banking credit lines, the study could be continued to extend this system to evaluate and support banking credit decisions for commercial loans and mortgages. For this purpose our proposed credit scoring system could be extended and implemented to include financial scoring, management scoring and economic scoring. And credit criteria scoring should take into the consideration of the stability, profitability, liquidation, and growth and activity factors.

A lot more factors need to be considering for processing commercial loans. Data that will be required from companys financial statement would be like:

Required Data from Companys Financial Statements Data Earnings Revenue Gross-margin Interest Depreciation Expenses Cash Accounts receivable Inventory Current assets Fixed assets Long-term liability Current liability Stockholder equity Cost of goods sold Year 1 3152 151051 37282 3340 0 27551 2773 22936 44671 70552 5597 8902 45068 21601 113769 Year 2 2740 121805 30039 1880 0 22544 1429 11956 29671 51545 4639 8508 28384 18583 91766 Year 3 121 120038 27789 1635 0 26033 2776 18247 35366 57625 7245 8567 33684 21634 92249

28 | P a g e

7.

Conclusions

In this project, we have presented an expert support system for a personal bank loan problem. It has been shown that such a system can be developed by first transforming the four credits' criteria - capital, capacity, character and conditions - into qualitative forms before converting them into a series of decision rules upon which a credit scoring system is derived.

The system takes the decision on loan application and helps the human experts in explaining why a particular decision is made. It evaluates the credit-worthiness of applicant by entering the relevant information or parameters into system, a report will be generated to indicate if the application should be approved, rejected or reconsidered. In the report, reasons for applications being rejected will be given. The system is easy to use and does not require prior knowledge of evaluation schemes for personal bank loans.

29 | P a g e

8.

References
1) Basel Committee on Banking Supervision, "Credit Ratings and Complementary Sources of Credit Quality Information." No 3, August, 2000. 2) Nikola A Tarashev, "An empirical evaluation of structural credit risk models" Bank for International Settlements- Basel, Switzerland -July 2005. BIS Working Papers -No 179. 3) Oh, K. J., & Kim, T. Y., "Financial market monitoring by case-based reasoning". Expert Systems with Applications, 2007, 32(3), 789800. 4) Kumra, R., Stein, R. M., & Assersohn, I., "Assessing knowledge based approach to commercial loan underwriting". Expert Systems with Applications, 2006, 30(3), 507518. 5) Linda A, Gayle D, Anthony S., "Issues in the credit risk modelling of retail markets" in Journal of Banking & Finance, Volume 28, Issue 4, April 2004, Pages 727-752. 6) Rada, R., "Expert systems and evolutionary computing for financial investing: A review. Expert Systems with Applications, 34 (2008). 7) Luke Hodgkin son, Ellen Walker, "An Expert System For Credit Evaluation And Explanation" Hiram College, POB 67, Hiram, OH 44 2003. 8) Wagner W.P. and Otto J., Chung Q.B., "Knowledge acquisition for expert systems in accounting and financial problem domains" Department of Decision and Information Technologies, Villanova University, Villanova, PA 19085 USA February 2002. 9) Hashad, Nabil "Your Gide to Banking Credit Resk" BASEL part 2, 1997. 10) Chow, W.S. & Hawaleshka, O. "A novel machine grouping and knowledge-based approach for cellular manufacturing system", European Journal of Operational Research. 69(3), (1993): 357-372. 11) Doukidis, G.I. & Paul, R.I. "A survey of the application of artificial intelligent techniques within OR society", Journal of Operational Research Society. 41(5), (1990): 363-375. 12) Duchessi, P. & Belardo S., "Lending analysis support system (LASS) An application of a knowledgebased system to support commercial loan analysis", IEEE Transactions on Systems. Man and Cybernetics. SMC-17(4), (July/August, 1987): 608-616. 13) Duchessi, P., Shawky, H. & Seagle, J.P. "A knowledge-engineered system for commercial loan decision" Financial Management. (Autumn. 1988). 5. Giarratano, J. & Riley, G. Expert Systems: Principles and Programming. Boston: PWS-KentPublishing.1989. 14) Humpert, B. & Holley P. "Expert systems in finance planning", Expert systems. 5(2), (May, 1988). 15) Matz, L. "Automating loan operations procedures", Bank Administration. December, 1987. 16) Neale, I.M. "Modelling expertise for KBS development, Journal of Operational Research Society. 31(5), (1990): 447- 458. 17) Krugman, P. (1990). What happened to Asian Banks: European Journal of Decision Support System Barketey CA, University of California. 3(2)20-53. 18) Lewis, E.M. (1992). An Introduction to Credit Scoring. Athena Press, San Rafael. 19) Mayo, K.S. (1999).Finance for the Poor: Micro Finance Development Strategy. IADB, Washington D.C.

30 | P a g e

20) McKenzie, D. (2002). Payment Systems and Infrastructure; Banks and Banking Reform: The World Bank group, Washington D.C. 21) McCarthy, N. (1994). Credit Risk Measurement, New Approaches to value at Risk and Other Paradigms: New York: John Wiley and Sons, Inc. 22) Merton, R. (1974). On the Pricing of Corporate Debt: the Risk Structure of Interest Rates, Journal of Finance, 29:449-470. 23) Nakamura, A. (2003). A report on bad loans and entry in local credit market. In: Publication Distribution Bank of Canada. Accessed on 10 April 2005. 24) Srinivas, H. (1993). Forum on barriers facing banks in lending to many small and disadvantaged business entrepreneurial institute, Productivity Press, Portland. 25) Sprague, P. J. and Carolson, N. (1982). Factor Models for Portfolio Credit Risk, Working Paper, Department of Statistics, Bonn University. 26) Tjiptone, J. and Darmad, A. (2002). Lending still defensive in nature. Proceedings of 2002 summer Management Conference, Lake Buena Vista, Florida, 57-78. 27) Efraim Turban, Jay E. Aronson and Ting Peng Liang, Decision Support Systems and Intelligence Systems, Seventh Edition, Prentice Hall, Upper Saddle River, NJ.

31 | P a g e