Research And Application Of E-commerce User Portraits Based On Big Data

With the rapid development of the Internet,data has become a necessary resource for the e-commerce industry.The capture and processing of data can no longer be done by conventional means.In the era of big data,the e-commerce industry is also facing greater opportunities.Among them,the collection of user data and personalized recommendations for products have become the top priorities of major e-commerce companies.For example,Alibaba and JD in the e-commerce industry use personalized recommendations for valuable marketing.Then,in the personalized recommendation system,the extraction of user portraits is the key to system implementation.With the increase of user data,the mined data models are gradually showing a diversified situation,which provides a wider space for the extraction of user portraits.This article uses data from an electrical appliance mall to propose a multi-dimensional tag such as rule matching,statistical,and data mining to form user portraits.The main points of this paper are as follows:1.Select the appropriate big data related framework and components for this system,form an overall framework structure,and set up the system's execution environment to lay the foundation for the function realization of each module.2.Import business data from MySQL into Hive,build a wide table,and then use BulkLoad technology to import the data into HBase.3.Use Spark to calculate the data to build a label model.The mining tags will be implemented using machine learning algorithms.This article uses Scala language to write a framework,implement functional programming,use Spark for a unified calculation processing analysis engine,use HBase to store user data labels,and finally implement labeling for users.