Design And Implementation Of The Intelligent Real-Time Transaction Risk Monitoring And Analysis System

With the booming of information technology and banking business, the E-bank, with its convenience, efficiency and low cost, has become a popular banking business and trading method. On the other side, there is potential risk on E-bank because it has been used through Internet and mobile network since people can use E-bank to trade on not only website but also their smart phone. People will easily lose their key identity information such as password if their guard consciousness of transaction security is not strong. E-bank fraud will not only cause client’s financial loss but also the bank image in client’s mind. Many commercial banks have taken some control and technology actions to prevent risks on E-bank. These actions such as password card, electronic scrambler, U shield, digital certificate, text verification, dynamic password, and etc. are quite useful to improve the prevention while decrease client’s satisfaction and convenience. Furthermore, some of these actions can be broken. For example, SMS Trojan can break text verification on smart phones. The new challenge is how commercial banks can reduce the loss on the convenience and efficiency as far as possible on the premise of guarantee transaction security.The thinking of improvement which this article mainly discussed about the intelligent and real-time transaction risk monitoring and analysis system is the path of transaction security protection developping toward real-time, intelligence and base on analysis of client’s behavior. The risk can be estimated by the comprehensive assessment combined the awareness of trading terminal environment, the transaction characters, nature and relevance of transaction, and transaction behavior and client’s behavior. Also, a real-time monitoring and analysis mechanism afterwards can be established to take appropriate safety measures of intervention to different types of transaction intelligently and timely. The complete solution is based on device fingerprint, Drools rule engine, Memcached, decision tree, Hadoop and etc.This article summarizes the research background and the course, introduction on related main technology concept, analysis of part of basic principles and causes of usage of the system first. Secondly, UML method is used to analyze the key requirements and the overall system is design into two subsystems. Then, the process of achievement on the online and offline subsystems is discussed through the aspects of system architecture, system interface, each system module, database, principles, and risk model design at the stage of design and implementation. In this article, the main work and improvable content as follows: · Analyzing business model and the flow on E-bank transaction risk monitoring and turning into system requirements;· Abstract and implement E-bank transactions fraud monitoring rules of business, flexibly configure mechanism and application on-demand; · Design and implement the algorithm and application based on fingerprint technology research, improve the invasive identification method on simplex using IP and MAC address of the terminal, improve identification on E-bank transaction terminal scientifically and accurately; · Achieve timely and efficient mechanisms and procedures including supporting statistical factor monitoring rules at a relatively low cost by using Drools and Memcached technology; · Generating reference rules and weights of characteristic variable by using machine learning on decision tree technology, improvement on the risk expert experience model, improvement on the scientific nature and accuracy of risk monitoring model; · Using Hadoop big data computing platform and customer behavior data to build the behavior model and applied to transaction monitoring. It improved the accuracy on recognition of individual trading risk.