Personalized recommendation method based on cloud computing processing mode in electronic commerce environment

Abstract

The invention discloses a personalized recommendation method based on a cloud computing processing mode in an electronic commerce environment and mainly aims to solve the problems that existing personalized recommendation methods are low in recommendation efficiency and poor in recommendation precision when processing mass data. The method is divided into an offline part and an online part, wherein in the offline part, a Hadoop framework of a cloud computing technology is adopted to parallelly process historical data information, a distributed file system HDFS is utilized to store mass data information, and four parallel recommendation methods suitable for different business stages of electronic commerce are implemented according to a MapReduce programming model; and in the online part, alightweight database is arranged to store user behavior logs, a dynamic data collection mechanism is designed to read data, obtained through processing in the offline part, in real time, and web display and information statistical analysis services are provided to provide real-time recommendation information for a user. The method has remarkable advantages when processing mass data generated by electronic commerce applications.

Description

Technical field [0001] The present invention relates to the field of computer network and data mining, specifically, it is a personalized recommendation method based on cloud computing processing mode for e-commerce environment. Aiming at the mass data characteristics and specific business logic of current e-commerce applications, it uses Hadoop framework to store Record user behavior and product information, and use Web logs online to dynamically record the user's current behavior and offline recommendation calculation results, so as to provide fast and efficient personalized recommendation services. Background technique [0002] E-commerce usually refers to a wide range of commercial trade activities around the world, in the open network environment of the Internet, based on browser / server applications, buyers and sellers conduct various business activities without meeting each other to realize consumers' online shopping, A new type of business operation model for online transa...

AI Technical Summary

Problems solved by technology

Personalized recommendation analysis for such a huge data set requires a large amount of computing and storage resources. If the centralized analysis and processing mode is still adopted, the recommendation time will be too long and cannot meet the real-time requirements, which will greatly affect the user's shopping experience; the data sparsity problem means that in a large-scale e-commerce environment, products rated by multiple users at the same time are very rare, thus affecting the mining accuracy of similar users; the cold start problem means that it is difficult for new users to calculate neighbors because they have no purchase records And new products are hard to recommend due to less ratings

In view of the above problems, the existing collaborative filtering recommendation methods are difficult to apply to the current e-commerce applications, so a new personalized recommendation method is urgently needed to solve the above problems and achieve fast and efficient recommendation goals

Images