Research And Implementation Of Marketing Platform Based On Collaborative Filtering Algorithm

With the popularity of the Internet,on-line marketing has developed rapidly.In order to help customers discover more potential users,increase the conversion rate of ads,and increase the revenue of customers,each internet company's advertising delivery platform provides a series of products,then it is particularly important how to effectively promote products to customers.For Baidu Inc,the existing product promotion model still has problems such as messy delivery platform,difficult maintenance,one-way promotion channel,long feedback path,low product conversion rate and coverage rate.In order to maintain Baidu's leading position in the advertising industry,it has become extremely urgent to solve these problems.First of all,this paper describes the background of the development of on-line marketing and the current problems of product promotion model in Baidu Inc,as well as the domestic and foreign development and research of on-line marketing and personalized recommendation,and further expounds the research objectives and research contents of this paper.Then,the relevant theories and techniques used in this paper are elaborated,including customer feedback,implicit user behavior data collection,Web data mining and personalized recommendation technology,which lays a foundation for further research.Then,this paper focuses on the personalized recommendation methods of the marketing platform,including data collection,data preprocessing,the “customer-product item” interest matrix model acquisition,and the generation of recommendation results.The data collection mainly studies the data types and collection methods of the marketing platform;data preprocessing mainly includes customer identification operations on buried log data,unrelated data cleaning,customer identification,session identification and path improvement for Web log data;The "customer-product item" interest matrix model generation mainly studies the construction and correction methods of the matrix model;the generation of the recommendation results mainly elaborates the principle of the recommendation algorithm selected in this paper.Next,this paper studied the design and implementation of the marketing platform,including requirement analysis,overall design,database design and implementation of each modules.At the same time of marketing platform research and implementation,requirements were fully verified and tested,and the test results were analyzed in detail.Finally,it summarizes all the work of this paper,and at the same time,some need to improve and further research direction are discussed.The marketing platform implemented in this paper has been officially launched after the test is passed.According to the actual operation situation and the corresponding statistical data,the marketing platform implemented in this paper solves the problems of product promotion model in Baidu Inc,gives the Baidu commercial platform the ability of self-marketing,and increases the amount of customers using the product,reduces the cost of promotion of Baidu Inc.This paper has completed the expected work well.