Design And Implementation Of Active Defense System Against Telecom Network Fraud For Campus

According to the requirements of the Ministry of Science and Technology in the special work points of 2020 03,"strive to create a batch of demonstration models in the aspects of 5G network integration of smart campus application and the new generation of information technology to boost student safety management",This thesis actively studies how to establish a comprehensive chain public network security active defense system in advance,during and after the telecom network fraud and other risky behaviors.Enhance the active security defense capability and level of the public network in the campus environment,so as to promote the faster and better development of 5G and at the same time enhance the active security defense system of the campus.The purpose of this thesis is to design and implement a aimed at college students group’s telecom network fraud active defense system,collecting students’ talk on the phone,information communication operators offer a large number of voice and data is used as the sample data pool,the algorithm such as speech recognition,text classification and building data model to call information to make a multi-angle analysis and accurate classification,to call the content of the information involved in telecom fraud of real-time identification and classification,then the result feedback in the system reference for schools and the police,finally achieve tracking of telecom fraud through intervention,the effect of early warning,And can carry on the precise positioning of teachers and students of anti-fraud education.In this thesis,we consult the related information about speech recognition,text classification and anti-telecom fraud,and ensure the technical feasibility of the whole system.In this thesis,we use Deep Speech2,an open source speech recognition framework of Baidu,and FastText,an open source text classifier of Facebook for fraud recognition.The experiments show that these two algorithms can meet the requirements of telecom fraud information recognition.After working out the system requirements with the school and the public security department,the system was divided into modular parts according to the system requirements.Finally,the overall construction of the system was realized based on the B/S system architecture.The system has been tested to meet the demands of schools and public security departments for telecom fraud,and has been put into trial operation in the Joint Research Center of Intelli Sense and Network Security of Nanchang University.Through a comprehensive analysis of the former lake campus of Nanchang University 35,000 students found that the number of students received telecommunications network fraud information is very large,comprehensive coverage of the department,the number of potential high-risk risk is a lot of people,the number of people at risk in the middle and above 3652 people.Through these data we can see that the school telecom network fraud development trend is very fast,the coverage of the population is particularly wide,the harm impact is very prominent.