Research On Intelligent Information Archive System Based On Big Data In C2B Mode

This article analyzed and explained the current limitations and deficiencies of the traditional user archive data information system based on the emerging e-commerce C2B(Customer to Business)mode.with the popularization of big data,this article needs to combine the large amount of archive information with the e-commerce model,to explore the information value of archives,and to achieve a certain degree of archive intelligent information.The main innovation points of this article can be summarized as the following three aspects:Firstly,proposed a hybrid recommendation algorithm that combined the user collaborative filtering and similarity algorithm to realize the distributed design on Hadoop platform.With the increase of the user,the acceleration ratio of the user collaborative filtering and the similarity will increase linearly,proving that distributed computer tasks can actively perform parallel processing without being constrained by performance,and this algorithm has considerable flexibility and scalability.Secondly,in addition to using Hadoop to implement the system prototype while solving the problem of massive data,this paper also used many practical open source tools such as Mahout,Sqoop,and ganglia to make the archive recommendation system have high-performance distributed parallel processing capabilities.Lastly,the study innovated the system mode,with the help of the C2B model,personalized customization can be realized to the greatest extent,while the most suitable product archive can be obtained through user information and product information.Therefore,the proposed methodology could greatly exert the true value of user archive.Rely on the above innovation points,this study constructed a general framework of intelligent information archive system based on big data in C2B mode,proposed a method to store the large amount of archive data in HDFS of Hadoop,and provided recommendation algorithm,which could parallel processing the distributed design and archive recommendation for Mapreduce.Based on this framework,the paper combined the archive data with the emerging e-commerce model,by using a hybrid algorithm that mixed the collaborative filtering algorithm with the similarity algorithm to maximizes the value of the user archive data and recommends the most suitable products to the user.