Design And Implementation Of Telecommunications Fraud Semantic Pattern Extraction System

In recent years,telecommunications fraud cases have become more frequent,the criminals;methods of committing crimes are becoming more and more sophisticated,and the communication content of telecom fraud is becoming more and more complex and changeable,resulting in huge losses to the people's property.However,because telecom fraud is still a small probability event in mass communication records,the proportion of fiaudulent content is low and the training samples are few.In addition,the criminals use synonyms,related words,and abnormal word order in communication content to evade telecommunication fraud screening,which leads to increasement of the difficulty of using traditional means to train and identify telecommunication fraud content.On the basis of fully considering the characteristics of telecom fraud,this paper proposes a telecom fraud semantic pattern model.The model builds a semantic model based on the scenarios and objectives of telecom fraud,and extracts the key information of the original communication content according to the telecom fraud semantic model,thereby removing redundant information,refining key content,which make telecom fraud content simple and clear.At the same t:ime,considering the small probability of telecom fraud content,the paper establishes a crude filter model of suspected fraud,and the initial filtering of the original communication content reduces the interference of normal communication content on the semantic pattern recognition of telecommunication fraud,and deals with the inefficiency of full content information processing caused by the low proportion of fraudulent content.According to the idea of semantic pattern model,this paper designs and implements the telecom fraud semantic pattern extraction system.This thesis first introduces background knowledge,as well as related techniques such as text classification and sequence labeling.Then it elaborates on the design of the main algorithms of the paper,including the crude filter model of suspected fraud and the semantic pattern extraction model.Through the analysis of the requirements of the semantic pattern extraction system,the various functional modules of the system are determined.Through the analysis and definition of the interaction process,interface and data structure of the system,the detailed design and implementation of each functional module are determined.Finally,through the system test cases and the analysis of the test results,it is determined that the system meets the functional requirements and expectations,which proves the feasibility of the system.