Research And Application Of K-means Algorithm Based On A-D Model

Voice calls include fixed telephone and mobile phone.It is a way of voice communication by transmission media.It is real-time,privacy and authenticity.It is also not limited by time and space.It is widely used information interaction mode.Although the voice calls bring great convenience to people's life,there are a lot of negative effects to a large number of customers and telecom operators themselves,such as telecom fraud,malicious arrears and advertising harassment.However,mining valuable information in a large number of call records requires a lot of manpower and material resources.Therefore,selecting effective algorithm by artificial intelligence and automatically mining valuable information from mass voice call records has become a research hotspot in recent years.This paper used the mass of call records to cluster analysis a variety of voice communication abnormal behavior,such as fraud customers,advertisers etc.The paper designed and constructed the model of abnormal behavior characteristics,based on which the model about a clustering speech communication algorithm for the analysis of abnormal behavior of customers was proposed.First of all,through the analysise of customer call records can got the customer calls,call connection rate behavior and so on.Then,combined the AHP model and DEMATEL method,construct the behavior model of customers' voice communication behavior.Secondly,an improved K-means algorithm was proposed based on the model,which realization clustering analysis of abnormal customers according to call records.Finally,the verification analysis is carried out by using real data.The results show that compared with other similar algorithms,the performance of this algorithm has been greatly improved in many types of abnormal customer integrated clustering analysis and single type abnormal customer clustering analysis.Finally,aiming at the practical application needs,we apply the research results to design and implement the related application system based on Hadoop.