Data Analysis Based User Profile Design And Application For Information Application

With the rapid development of mobile Internet and the continuous popularization of mobile devices,mobile devices are becoming the main source of information for Chinese people.After the Internet has gradually entered the era of big data,it inevitably brings a series of changes and reshaping to enterprise and consumer behavior.One of the biggest changes is that all the consumer behavior seems to be "visualization" in front of the enterprise.With the in-depth research and application of big data technology,the enterprises increasingly focuses on how to use big data for accurate marketing services,and further tap the potential business value.In this thesis,information application is used as a breakthrough point to construct user profile.This thesis is from the actual business needs of a well-known information application enterprise,and it aims to practice the concept of precision marketing by enhancing the user activity.It needs to construct user profiles of the system based on user attributes and behaviors,which will be used as the basis of precision marketing and feedback.The main contents of this thesis include three aspects: analysis and division of the user activity problem,research and design of user tags and portraits,data mining of the active user profile and tags.The final output is in the form of a user's tag library and a user's portrait library.The main work includes:(1)Analysis and division of the user activity problem.This part studies the two dimensions of the breadth and depth of the user activity.It determines the design direction of user data labels.(2)The design of user data tag library and the collation of basic data.This work is to classify the user information based on the current online data.(3)Cluster analysis and user profile design.This part of the research uses clustering analysis techniques to classify users based on the user data tag library and designs business portraits based on users of different characteristics.(4)Research on Data Mining Based on activeness association.In this part,the association rule algorithm is used to mine the profiles related to activeness,and the degree of association is quantified.(5)Evaluating the effect of mining results in practical application.This part of the work is to verify the effect of the application level of the portrait,and use the current popular AB test method to evaluate the effect quantitatively.From the final on-line experiment results,the validity of the portrait system can be verified.The clustering method and the research methodology of data mining can effectively improve the user activity.At the same time,the establishment of the whole profile system for enterprises plays a positive role on optimization of the user experience,the commercial competitiveness and integration of market resources,and laid a solid foundation for better service to customers and achievement of the business value.