The Research And Design Of Big Data Recommender System Between Enterprises

In recent years,with the continuous application of big data thinking in various fields of industry,the huge value of the data gradually revealed.Many major companies are competing to build their own big data platform to collect,mining and analyze data,to make business developed in a more scientific,more efficient,more intelligent direction.However,due to the huge difference between business and data,how to effectively collect the stock of big data and mining value of big data becomes a challenging application research topic.The research and development of this project come from the real application requirements of a group company,which is based on the application of the model and algorithm design of the big data recommendation between enterprises.On the basis of big data platform,the system of this project profiles for user with proposed tag segmentation algorithm,profiles for target group users of merchandise with designed feature tag extraction algorithm,implements mutual recommendation of customer and merchandise between enterprises based on profile data matching.The main research contents of this paper include:(1)This paper proposes an algorithm RR-SEG,which is a rough and refined tag segmentation algorithm.It groups and counts data after sorting,and computes the edge data of each segment in order according to average count of every segment,to implement uniform segmentation for different distributed data,ensure objective and exact of segmentation result.(2)Designs a proper approach for feature tag extractio n.It computes the variable coefficient of standard deviation according to the count of each tag value,uses the variable coefficient as measure index of dispersion,sets as feature tags with those tags whose variable coefficient exceeds the threshold.Feature tags is able to exactly reflect the influence bigger factors of merchandise purchases,it's an important method to analyze the feature of merchandise target group users.(3)Designs and implements a recommender system between enterprises.Under the big data system architecture,and based on user profile and merchandise profile,the system matches tags of merchandise target group users of objective enterprise against user tags of other enterprises,to achieve mutual recommendation of customers and merchandises between different enterprises.On the whole,the paper proposes the enterprise big data recommender system which can collect members and sales data between enterprises effectively,recommend potential members and related merchandises that meet t heir requirements for various enterprises with the data.System test results show that this system can effectively help enterprises get new customers,improve the conversion rate and enhance the performance of enterprises.