Building a Rule-Based Credit Risk Assessment Expert System - Research Paper Example

Building a Rule Based Credit Risk Assessment Expert System Explain the Importance of Credit Risk Assessment The credit risk assessment or analysis and credit risk management are vital practices for the financial institutions that provide a loan or credit services to individuals and businesses. Normally, credits can be offered to individuals or businesses to facilitate repayment of their bank mortgages, to purchase vehicles, credit card purchases, and installment purchases among other financial usage. In some cases, credit users may default repayment of these credit and loans services (Hayre 56). Therefore, lending institutions normally deploy numerous and varied techniques of credit risk management to understand levels of risks associated with credit users (Grzymala-Busse 25). In most cases, statistical preventive analytical techniques are used to analyze and determine credit risk levels associated with loans and credits given to borrowers. Therefore, the significance of credit risk assessment is to reduce credit or loan defaulters. Adequate information about the credit users are often for the analysis of credit risk levels. This information is usually obtained from internal credit scoring systems. This system allows computation of personal information credit scores from credit reports (Camp 14). Such information is provided for by rating agencies or external credit bureaus. Notably, the credit scores often indicate an individual or organizations’ current and historical financial situation. These financial reports are then used to analyzing credit risk levels thereby determine credit defaulters (Grzymala-Busse 71). However, the internal credit scoring techniques do not define defective or ineffective score. Therefore, it does not predict the actual levels of risk associated with lending a person. The shortcomings of the internal credit scoring methods have been solved by the use of the Profiling risky credit segments. This method is tremendously significant in assessing credit risk levels. It applies The Pareto principle that suggests that the majority (eighty to ninety percent) of the credit defaulters emanate from lower (ten to twenty percent) lending segments. Therefore, segment profiling usually provide vital information for credit risk analysis. In this analysis, Credit providers usually collect vast credit information or data of the credit users (Graham and Milne 38). Numerical and categorical data concerning the credit users are collected. Since the collected information is never synchronized, it is often considered noisy or insufficient. Therefore, profiling facilitates the identification of variables or factors that accurately summarizes the segments to give sufficient information for analysis of the credit risk levels (Zopounidis and Doumpos 19). Other than profiling, accuracy of the data collected about the credit users can be guaranteed by use of the Hotspot profiling analysis. This analysis allows drilling of data systematically to detect accurately the vital relationship, interaction, cofactors, association, and dependency of the provided data (Zopounidis and Doumpos 82). The drilling data analysis uses the artificial intelligence techniques in generating profiles for the targeted segments. Therefore, the hotspot analysis provides accurate identification of low and high loan or credit risk levels or defaulters. Moreover, credit risk modeling techniques have been initiated to enhance assessments of risk levels. These techniques provide guild of predicting future events. The predictive modeling techniques are perfect ways of managing credit risks (Yu 29). The predictive models help in developing historical records of financial, credit loan, psychographic, demographic, and geographic information concerning the credit users. From the historical credit information of an individual, the predictive models can predict a different pattern in the credit default ratio thereby predicting future credit and loans risk levels associated with the credit user. Notably, this statistical method provides vast, substantial historical information or record concerning the credit users; thus, facilitating credit risk level management. Explain Rule Based Expert System Despite the use of these models to predict credit defaulters, Real time experts warn that lack of historical, statistical evidences may hinder the effectiveness or reliability of these predictive models (Camp 59). Therefore, they advise that the credit risk assessment for the high-risk applications to be filtered manually using various policies, regulations, and judgments using rule based expert systems. They further highlight that, the high-risk credit statistical evidences often lack in the historical data; thus, the use of the rule base system allows a paradigm knowledge representation. This system applies the use of knowledge contained in a knowledge domain code that is executed in the form of rules. The application of this system highlights or predicts high risks that are unlikely to be predicted by predictive models. In other words, the credit assessment cannot only apply the use of the predictive modeling (Graham, and Milne 275) but also requires knowledge or rule based expert system for effective credit levels assessments. Therefore, an effective credit assessment employs the use of rule based modeling and experiences of human experts. The rule-based expert system also uses Rosella BI Platform that applies the two rule-based modeling techniques including RME-EP and RME both of which are based on SQL language specifications. They form immensely powerful languages that are used in predictive models alongside mathematical formulas and logical expressions (Yu 51). The RME is used as a procedural language where RME-PE applies for the rule-based experts systems. These two languages collectively form an extremely powerful risk-modeling platform. The rule-based expert system uses software that has a user interface that supports the communication between the system and the user. Through the user interface, the lender keeps a track that presents the chain of reasoning that leads to a specific conclusion. The features of system expert have vast difference from other conventional systems. The system’s explanation facility enables the commercial experts to carry out trace-based explanation depending on the specific input data set. The expert system also has working memory that collects and stores information that are later used as rules. These rules are what facilitate credit risk assessment. Working memory and inference engine often access and match facts with high precision thereby being used in baking systems to predict credit defaulter. The inference engine provides satisfied facts that are executed in two ways including forward and backward chaining. The forward chaining provides factual reasoning that leads to a conclusion while the backward chaining provides hypotheses that support these facts. Therefore, the result or conclusions arrived at, using expert system, depends of the input data and the problem under investigation. In the banking system, the backward chaining is often used to diagnose the problem while the forward chaining is used for monitoring, prognosis, and controlling the risk. Therefore, the rule-based expert system facilitates explanation building (in relation to borrower’s available data) during credit risk assessment in the banking system. Example of Applied Rules in Credit Assessment (Mortgage) There are numerous rules applied during the process of credit assessment. For instance, let us consider credit risk assessment in mortgaging. The initial step in mortgaging is the need of buying a property. In this case, professional team will provide an explanation for the processes involved towards possessing the mortgage or the buyer’s new property (Yu 88). After the completion of the preapproved mortgage application, the mortgage broker calculates the maximum purchase price the buyer can afford. This is done using the information or data obtained from buyer. The legibility of the buyer obtaining the mortgage will depend on the analysis of the set rules by the mortgage lender. The first rule examines the buyer’s Gross Debt Service (GDS). This rule makes the buyer not to remit monthly housing or mortgage cost that exceeds 32 percent of her or his monthly gross income. In other words, the housing costs should always include the buyers’ monthly mortgage payment, property taxes, and heating amount. However, purchasing strata property will include half strata fees (Hayre 72). The Gross Debt Service is usually calculated by dividing the total monthly payments by the gross monthly income. The rule of Total Debt Service (TDS) states that, payment of monthly debt should not exceed forty percent of one’s gross monthly income. The TDS includes the housing costs and other monthly payments (Angelov 97). The Total Debt service is obtained by dividing the total monthly debts by gross monthly income gives the (Abrahams and Zhang 312). The other sixty percent of income is expected to be used in other monthly expenses including income tax, hydro bills, phone bills, entertainment, and vehicle or house insurance among other expenses. Notably, the above two rules are often ignored by some leaders when seeking for down payment from the credit users. The third rule is often examines the borrower’s Credit Rating. In this case, the loan officers seek the permission of the credit user to obtain his or her credit bureau credit report. This report shows current and previous credit ratings of the credit user. The credit users who often slop their payments often have negative credit ratings. Personal credit history usually determines difficulties associated with the loan or credit approval (Angelov 201). The lenders often determine the effectiveness of this rule. However, the credit bureau is an essential record in determining how one handles his or her credits. Individuals with higher rates stand risks of being denied credit. It is only after considering these rules that the credit providers will determine whether to offer credit or not. Regulations of Basel II and III Basel regulations are international continual recommendations that aim at stabilizing banking systems even in times of management and financial crises. The Basel II defines the regulations of the Basel Second Accords that have been extended and effectively superseded by the Basel III. The Basel II recommended the banking laws and regulations should be used in supervising banking systems (Mira 274). It intended to create a regulation that would have initiated the international standard that could regulate the capitals that banks should set aside against any financial and operation risks that may face the banking system due shift in the global economy. This initiative was geared to rescue banks from any collapse. However, the global politics hindered the implementation of the Basel II recommendations (Tarullo 172). Failure of implementation of Basel II led to the banking crisis that was mainly caused by mortgage backed security, credit default swaps, and other similar derivatives. This banking crisis led to the establishment of Basel III. The Basel III recommended comprehensive set of measures or reforms. Notably, the recommendations of Basel III strengthened the supervision, regulation, and risk management within the banking sector. These reforms aimed at improving the ability of the banking sector to absorption shock towards economic and financial stresses thereby improving governance and risk management as well as strengthening the disclosures and transparency within banks (Tarullo 28). The Basel III also introduced regulations that would affect banks at certain levels or the micro prudential as well as raising their resilience of distinct banking institutions within the stress period. Finally, Basel III defined the macro prudential systems across banking sectors. Furthermore, its procyclical amplification were to apply whenever the risk has gone overtime (Tarullo 201). Both Basel II & III regulations are continuous efforts of the Basel Banking Committee that aimed at enhancing the regulatory framework in the banking system. Additionally, they enhance international convergence of capital standards and capital measurement of documents. Works Cited Abrahams, Clark R, and Mingyuan Zhang. Credit Risk Assessment: The New Lending System for Borrowers, Lenders, and Investors. Hoboken, NJ: Wiley, 2009. Print. Angelov, Plamen P. Evolving Rule Based Models: A Tool for Design of Flexible Adaptive Systems. Heidelberg [u.a.: Physica-Verl, 2002. Print. Camp, Olivier. Enterprise Information System V. Boston: Kluwer Academic, 2004. Print. Hayre, Lakhbir. Salomon Smith Barney Guide to Mortgage-Backed and Asset-Backed Securities. New York: John Wiley & Sons, 2002. Internet resource. Graham, Ian, and Robert Milne. Research and Development in Expert Systems Viii: Proceedings of Expert Systems 91, the Eleventh Annual Technical Conference of the British Computer Society Specialist Group on Expert Systems, London, September 1991. Cambridge: Cambridge University Press on behalf of the British Computer Society, 1991. Print. Grzymala-Busse, Jerzy W. Managing Uncertainty in Expert Systems. Boston [u.a.: Kluwer Acad. Publ, 1991. Print. Mira, J. 11th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, Iea-98-Aie, Benicàssim, Castellón, Spain, June 1-4, 1998. Berlin: Springer, 1998. Print. Tarullo, Daniel K. Banking on Basel: The Future of International Financial Regulation. Washington, DC: Peterson Institute for International Economics, 2008. Print. Yu, Lean. Bio-inspired Credit Risk Analysis: Computational Intelligence with Support Vector Machines. Berlin: Springer Verlag, 2008. Internet resource. Zopounidis, Constantin, and Michael Doumpos. Intelligent Decision Aiding Systems Based on Multiple Criteria for Financial Engineering. Dordrecht [u.a.: Kluwer Academic Publishers, 2000. Print. Read More