Design And Implementation Of A Telecom User Abnormal Behavior Detection System Based On Communication Social Behavior Learning

In recent years,with the development of communication technology and the popularization of mobile devices,communication behaviors such as telecommunication fraud and malicious harassment have greatly affected the order of society.In order to detect and intercept such abnormal behaviors of telecom users,existing methods use statistical learning and machine learning techniques to model and learn data by constructing user communication behavior characteristics and social topology information to detect and identify abnormal users.However,these methods need to collect long-term user communication records to construct user samples,resulting in a lag in anomaly detection.In order to improve the timeliness of the abnormal behavior detection task of telecom users,this paper designs an anomaly detection method based on communication social behavior learning,considering the characteristics of user communication behavior and the evolution characteristics of communication social network,aiming at the short-term communication behavior data of users.This method introduces a local community discovery algorithm to generate the user’s communication social network,effectively extracts the communication social relationship,and reduces the noise data in the communication network.After that,a communication social behavior dynamic network is constructed on the basis of the user’s short-term communication session sequence with introducing a long-short-term memory to evolutionary graph convolution model to learn the social nature of communication behaviors,and then realize anomaly detection of user communication behaviors.The comparative experimental results show that the method proposed in this paper can improve the accuracy and precision of anomaly detection under the limited amount of user communication record data.On the basis of the algorithm,the paper designs and implements a telecom user abnormal behavior detection system based on communication social behavior learning.The paper first analyzes the requirements of the system,then proposes a communication social network generation scheme and a dynamic network modeling method for communication social behavior,and describes the evolutionary graph convolution model based on communication social behavior learning in detail.Then,according to the requirement analysis,the general design and detailed design of the system was completed,the anomaly detection algorithm and system function modules were implemented,and the system implementation effect is displayed.Finally,this paper designs detailed test cases for the functions of the system,and conducts functional tests and performance tests to verify the correctness and usability of the system.