With the increasing development of the financial markets, fraudulent activities is severely growing in state-owned commercial banks (here, commercial Bank is the commercial in nature). So the state-owned commercial banks are facing greatly increased operational risks. As the structure of State-owned banks has changed greatly and financial business process is continuously being reformed, a sound risk controlling system and the corresponding internal controlling system must be established in order to effectively prevent risks and avoid risks.In this paper, Bank Fraud Screening Platform is a part of the Bank Audit System. So, we label it as the Commercial Bank Fraud Screening System. Simply, the system monitors the legitimacy and the regularity of banking transactions and the regularity of a clerk's operating. The system is an application module designed for commercial bank auditing organ and auditors. The aim is that filtering doubtful transaction data from transaction data real-time and warning doubtful trading operations according to the data item and the manual configuration in the alerting rules of system. On the premise of electronic business system fundamental, we use advanced computer processing technology to increase the automation level of monitoring transaction risk, to strengthen the investigation management of warning data for clerks, and to enhance the operational efficiency of clerks.As we all know, the existed ways of fraud screening mainly depends on database technology to complete the monitoring of transaction data through stored procedures and embedded SQL applications. There exists a problem that it's difficult to maintain and it's not efficient to solve problems. And the traditional development process model is"program = data structures + algorithms", but algorithms is exported difficultly in the case of complex business rules. And the main processes of software project development is demand analysis,summary design,detail design,encoding,testing. However business rules of market demand often change and business rules often embed into the code in the system. So design and development of the system is affected and the rules are updated. Due to the characteristics of fraud screening applications - the massive data environment, that monitoring of complex rules is very useful. So we put forward the design and implementation of the commercial bank fraud screening system based on the rule engine technology.Rule engine is a kind of inference engine, and born of RBES (Rule-based Expert System). Rule engine can be seen as a set of software components, which is responsible for extracting business logic from the application programs, and using a prepared semantics to edit the business rule into business decisions. It is an embedded component in the application. It is used to test and compare of the business rules loaded to the engine and the current data submitted to the engine, to activate the business rules connected to the current data, And to trigger the operation of the corresponding application according to the statement in the implementation of business rules logic. When we need apply complex rules on large amounts of data, rule engine technology is particularly useful. As a result, this paper presents design and implementation of fraud screening subsystem based on rule engine. In practice, I try to do the following innovations:1. Designing the rule engine and implementing rule file: It implements rule processing engine based Drools(JBoss Rules) and rules management. According to rule language-drl and development experience, I define a usable rule frame.2. Maintenance rules: I offer a interface for editing rule, so the clerks of bank can assign specific staffer to edit new rules based on requirement.3. Systemic efficiency: Solve the question of systemic efficiency in filtering massive amounts of data.4. Warning system: Build up the warning system with Chinese bank characteristics according to requirements of CBRC (China Banking Regulatory Commission).5. Effectively integrate the foreground editing rules, the background getting warning information by calling rules, So the system is more perfect and operation is more convenient.In this paper, I introduce theoretical knowledge of the system—rule engine, describe general framework of system,background design process, illustrate the design of processing background rule and displaying foreground interfaces and display partial results in detail. In the whole development process, the task of this paper has editing background parts rule files,designing the system including the whole design and the foreground design,developing foreground parts modules. Summarized as follows:1. I generalize the backgrounds,characteristics and working principle of rule engine, analyze structure,working mechanism,pattern matching algorithm of Drools rule engine. I study the working principle of Rete Algorithm and editing framework of rule file in rule language-drl. According to the background of the system, we determine the strategy and methods of applying rule engine. In the light of rule language-drl and development experience, I define a usable rule frame.2. Based on the risk management rules that Basel Banking Regulatory Commission, China Banking Regulatory Commission and other organizations put forward, China's major commercial banks build own rule sets in the light of own requirements. In this paper, we build up a complete banks risk warning system with Chinese bank characteristics according to the client's requirements.3. I offer a interface for operating esaily. So when the bank has new rule demand or updates the existing rule, the clerks of bank can assign specific staffer to edit new rules based on requirement and to maintance it.4. According to the bank requiremeng, we effectively integrate the foreground editing rules,the background getting warning information by calling rules,foreground displaying warning information and a series of investigation result for warning information. So the system is more perfect and Operation is more convenient.5. In the background data extraction process, we cite the Drools rules engine to solve the efficiency filtering questionable data from massive data. The system that completes calculating warning information daily reduces the working hours of clerk deeply. Due to pertinence that is strong,information that is comprehensive,finding program that is more and precise, the system improves the efficiency of the clerk significantly in the bank's testing and application.This fraud screening subsystem has been applied to accurately in a commercial bank, and efficiently to assist operations staff completed the data collection and filtering problems, which greatly improved the efficiency and quality of technical personnel. The client said that the system brought fraud screening to new life, became an indispensable tool and commended by the lead of the bank. In the relevant report they have repeatedly done instructions that they confirm the result of the system. |