Design And Research Of User Portrait System Based On Big Data Technology

With the continuous development and popularization of the Internet,the number of people using the Internet has been rising year by year.The behavior data left by users in the network is growing rapidly because of amounts of information.The accumulation of huge amounts of data led to the inundation of effective information,which not only made it difficult for users to obtain valuable information,but also led to the reduction of the conversion rate of value users in the enterprise platform.Therefore,it is very important to build a user portrait system that can help enterprises to understand users,locate users and tap users’ features.The core goal of this dissertation is to research and design a scheme for constructing user portrait system based on the existing big data technology.The main work includes:researching how to build current user portraits;researching how to effectively apply and support the construction of user portraits in current big data technology;taking"Heart Discovery" as an example,building a user portrait system which is fit for PGC(Professional Generated Content)platform.The process of building a user portrait system includes:defining the business needs of a user portrait,constructing a multi-dimensional user portrait labeling system,establishing matching rules for user portrait labels according to different business scenarios,establishing group portraits for users with similar user portraits and establishing effective labels update mechanism.In this paper,the system implementation scheme is proposed:the big data technology is used to collect and store the original data to provide the system performance support;the data structure and storage method of user portrait and multiple tagging systems are introduced;the user portrait label in the logical hierarchy between the conversion relations,and the formation process.Finally,the effectiveness of the user portrait system is verified experimentally.