Development And Implementation Of Decision Engine System For Internet Finance

With the rapid development of the Internet finance field,consumer finance microfinance for ordinary users has developed rapidly to adapt to market demand,and has become a new development driver and important financial innovation in the current economic field.Not only the extremely rapid growth of user transaction data and frequent changes in business scenarios,but also the diversified,intelligent and time-changing risks faced by Internet finance companies are required to quickly identify risks,improve risk management capabilities,and adjust business rules in a timely manner.It has become the most important and urgent issue facing companies and financial institutions.In the early stage,the risk control logic was deployed in the business logic through hard-coded rules,which led to problems such as strong coupling between rules and business,scattered and unmanageable,and long strategy cycles.It is proposed to use the decision engine to realize the decoupling of rules and business systems,the risk changes quickly,the configuration method of policy rules is simplified,and the verification and deployment time period is shortened.This article mainly studies the decision engine system for Internet finance,the main content includes:decision engine rule management subsystem and rule execution subsystem.The research content of this article can be divided into the following parts:First,analyze the challenges and project background faced by Internet finance,and introduce the rule engine in detail,including the rule engine algorithm and the open source rule engine framework.Secondly,design the overall scheme of the decision engine system.According to the usage of multiple business lines and business scenarios,the overall design of the user's business data isolation and authority management in the decision-making rule management subsystem is carried out.Considering the diversity of expressions of rules,a variety of rule configuration modules are designed,including general rules,rule tables,rule trees,score cards,and rule streams.At the same time,version management of decision-making releases is carried out to support online multi-version operation and rapid rollback.Business analysts need to design query backtracking rules execution details for the execution of decision rules.According to the analysis of system requirements,complete the outline and logic design of the database.And detailed design space project isolation business,user and authority management all user data and functional authority,support the principle of least authority.Design feature and rule management module,decision release and version management module.HTTP is used as the internal and external communication protocol of the system,Maven is used as the storage center for rule packaging,and Kafka is used to subscribe to the rule execution log and is subscribed to analyze and store it to the ElasticSearch search engine.Finally,the interface for rule execution and rule service management is designed,and the realization effect of the decision engine system is demonstrated through the credit scenario of Internet finance.The function and performance of the decision engine system are tested.The test results show that the system runs well and meets the basic design requirements.