Research And Implementation Of Bad-call Prevention Subsystem For Telecommunication Network

In recent years,criminals have used phones to defraud,and spread the information such as terrorism,prostitution and cults.Every year people lose tens of billions of dollars due to bad calls.Finding effective solutions to protect people's lives and property is urgent.However,at present,the bad call can only be prevented by raising people's awareness of prevention,and there are no effective technical means to prevent the occurrence of such incidents.Therefore,how to effectively detect,identify and deal with bad calls is called an urgent priority.This paper designs a set of bad call prevention system for telecommunication network,which can detect,identify and deal with bad calls at the communication network level.This paper mainly contains the following contents:first,the method of implementing telephone interception.This system is mainly oriented to IMS network,which uses the characteristics of the SIP protocol and constructs the BYE or CANCEL message to intercept the bad calls.Second,research how to find bad call numbers.Using the feature engineering found the embedded rule of call records,and creatively put forward the model which can find the bad call number.Improve the capability of the system to prevent bad call;Third,research the method of analyzing the call voice to find the bad call.Combine the convolutional neural network,Long Short-Term Memory and attention model to put forward AC-LSTM model,and the model compared with other models in identifying the bad calls has the best effect.Finally,design test cases to test the system's function,performance and recognition accuracy.The test results show that the system achieves the goal of accurately identifying and intercepting the bad call.