Research On The Model For Fraud Detection Of Campus Card

It is an accomplished fact that several research groups have retrieved the algorithm and developed attacks to break keys of MIFARE Classic-enabled cards, in 2007. System integrators therefore have to reconsider whether they have implemented appropriate security measures for the use of the MIFARE Classic card for applications that need security. This type of card was selected for most contactless campus card projects in china, The incident has caused us to think, we found more about campus card fraud. Campus card fraud would disrupt the order in school, and cause great losses to university, businesses and the cardholder, Therefore, the paper is of great importance not only the theoretical significance but also the practical value.Fraud Detection of Campus Card is defined as using some methods to check the abnormal transaction and assess the risk of fraudulent transactions when the fraud happened. It was designed to help improve the efficiency of administrator monitoring abnormal transaction, in order to achieve the purpose of increasing the safety of the capital which in the campus card system.In order to get rid of the capital safety problem resulted from the fraud of the campus card. This paper proposes a procedure of fraud detection named "Reconciliation-pretreatment-detection with neural network" which was based on the full research of the campus card system. A new algorithm is designed to detect campus cards which have abnormity transactions and then the scope of the campus card fraud detection could be reduced largely through it. A rule library that applies to campus card fraud detection was established also, and a model for fraud detection of campus card is built. The work mainly includes the following aspects in this thesis:1. A data set for campus card fraud detection was established through analysis, screening and treatment the real-world data that included in campus card system.2、For getting the characteristic information what needed in the campus card fraud detection, the historical transactions was analysised and mined through mining algorithm based on BP neural network in this paper. 3、we get the appropriate rule for the fraud Detection of campus card based on the fraudulent transactions, and then build up the neural network to detect the fraud of campus card. Finally, the model of the detection system for fraud campus card will be constructed.4、Experiments was designed to analysis and verification the effectiveness of the campus card fraud detection model.It proves that the work for fraud Detection of campus card was meaningful and feasible. It is helpful to control the bookkeeping risk which possibly exists in university, the conceal risk was defined and controllable; At the same time, from passive reporting to the initiative to find the fraud of campus card, which helps to shorten the discovery time of the campus card fraud and reduce the loss of cardholders and merchants.