Font Size: a A A

Design And Implementation Of Semantic Classification System For Telecommunication Fraud

Posted on:2020-04-18Degree:MasterType:Thesis
Country:ChinaCandidate:L M PuFull Text:PDF
GTID:2416330572473615Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the increasing number of telecommunication fraud cases in recent years,intercepting fraudulent telephones has become a key part of telecommunications security protection.Most existing fraudulent phone identification methods rely on data such as fraudulent phonenumber libraries,fraudulent voice templates,and user phonecall records.However,these methods are difficult to deal with the volatility of the means of fraud and language volatility.In order to solve the above problems,this paper designs a fraud semantic classification algorithm based on natural language processing related technology and communication fraud semantic pattern.The algorithm can analyze the call content of the incoming call at the semantic level,and calculate the fraudulentity of the call content and the semantic similarity between the content and other fraudulent call content.Content fraud is combined with fraud and similarity to deal with the volatility of telecom fraud.Firstly,the paper expounds the background knowledge of telecom fraud and the related techniques of fraud semantic classification.Then,based on the problem of language volatility and the complexity of semantic pattern structure,this paper proposes a hierarchical construction fraud semantic algorithm,mapping semantic patterns into a semantic vector space,and these fraud semantic vectors are used for fraud determination.Afterwards,the paper analyzes the telecom fraud semantic classification system requirements in detail,then designs and implements the fraud semantic classification system based on the fraud semantic classification algorithm,and the key implementation process is elaborated.Finally,the paper designs related test cases to test the system,and compares the fraud semantic classification algorithm to verify the effectiveness of the telecom fraud semantic classification system.
Keywords/Search Tags:telecommunication fraud, semantic classification, content fraudulent, natural language processing
PDF Full Text Request
Related items