Research On User Portrait And Tag Library Based On Zhujiang Big Data Project

With the rapid update of modern media technology,As a traditional media business,television is constantly being hit.The data show that since 2014,the market share of China's television broadcasting and television viewing has been analyzed,and the share of the broadcast is falling,and the negative growth of the ratings in many local stations has been continued.One of the main problems facing the TV industry is how to better understand the customer's needs,draw a precise user portrait,save the lost customers,provide the accurate service for the customers and realize the transformation of the traditional TV market.Based on the users' data of the "Zhu Jiang" group,this paper establishes a user tag library and draws a precise user portrait.It realizes the retention of the lost customers and lays the foundation for the subsequent intelligent recommendation.This paper first divides the various ways of the label generation into four categories,in which a customer is extracted from a real estate.According to the user's behavior data in a period of time,the user value model is constructed by K-means algorithm,an accurate label is attached to the customer,and the multiple behavior index of the user is analyzed,and the users of different classes are analyzed.People's preferences for TV needs and habits are the two most important ways.Taking a real estate customer data as an example,the paper divides customers into 6 categories through analysis,and gives corresponding retention strategies and marketing suggestions based on different types.Secondly,by tagging the data,we can improve the tag library,so that it can't be omitted,and it will be exhausted.Finally,the user portrait is actually a mathematical modeling of the real world users.The establishment of tag system is the core of user portrait.Tag is a symbol representation of user characteristics,and user portrait is represented by a set of tags.However,unlike the existing label system,"Pearl River digital" is mainly oriented to group customers because of its industry characteristics.Therefore,the label system is a group label system,and the same user portrait is the portrait of a group user.Compared with the individual portrait,the group portrait is more difficult to be accurately and accurately described.Based on the characteristics of the "Zhu Jiang" industry,this paper builds a complete group customer label system and draws a group customer portrait,focusing on the problems of the difficulty in constructing the group customer label library and the definition of the customer's portrait.According to the matching labels of customer portrait and product portrait,we can transform the marketing mode and provide more intimate and meticulous service for our customers.At the same time,it also provides reference for reasonable matching products.