Research And Application Of User Behavior Pattern Mining In Mobile Internet

The rapid development of Mobile Internet has penetrated into every aspect of people's life,and has become an indispensable element of human life in the information age.The user is connected to the mobile Internet at anytime and anywhere through a smart phone,and a large number of data is generated through the Mobile Internet.The user behavior patterns in the data are both opportunities and challenges for the development of mobile operators and human information society.Behavior patterns of mobile Internet users represent a summary of user behavior,which can be used to build the profile of user,find user's preferences,make user prediction and recommendation,and improve the service provided by the merchant.In order to mine the user's behavior pattern and analyze the user's behavior in detail,this thesis focus on studying and applying the behavior patterns of Mobile Internet users,including two main aspects:the user's click pattern mining and the user's mobile pattern mining.Firstly,this thesis introduces the background and significance of the Mobile Internet user behavior pattern mining.Secondly,we describe the mobile Internet user behavior pattern mining method,and focuse on the application of the pattern mining method in the user's click pattern mining and mobile pattern mining.Thirdly,this thesis introduces the technology of massive data mining based on cloud computing,including the distributed file system HDFS and distributed programming framework MapReduce in the distributed processing framework Hadoop.After that,we introduce the principle and method of the pattern mining algorithm based on Hadoop platform.Then,the user clicks pattern mining and user mobile pattern mining are described in detail.We mainly introduce the process of the click pattern mining and the mobile pattern mining.Furthermore,we analyze the results of the pattern mining and mobile pattern mining.The mining results of click patterns are applied to the analysis of the quality of the web page.The mining results of the mobility pattern are applied to the user's mobility prediction.Finally,this thesis summarizes the results of the mobile Internet user behavior pattern mining.