| With the boom of "Internet+",to satisfy the demand for enterprises' informatization and decrease the cost of enterprises' informatization,industry application gradually receive attention from people.The development of industry application products is rapid,and a growing number of industry applications are published on the industry application platform.Nevertheless,on this platform,there are a lot of similar applications in same field,such as ERP,CRM and OA etc.Simultaneously,all of the enterprise customers,application developers and operations personnel are facing with the challenge.For enterprise customers,how to find the ones needed from all applications will become increasingly difficult.For application developers,how to make their applications stand out and got the attention from the enterprises is also a very difficult thing.For operating personnel,how to make the product marketing target the user precisely,becomes the core of the operation.Therefore,the study of personalized recommendation system is of great significance.This paper base on Biznavigator platform's industry application recommended requirement,analysis the Biznavigator platform existing business content and business flow,researching on the user operation behavior of the application,found that the correlation between applications,and then based on these correlations to find suitable for industry application of recommendation method,and take advantage of this method design personalized recommendation system to industrial application for enterprises.This system build on the Hadoop open source framework,using Hadoop,create an independent,reliable,efficient and scalable system,using collaborative filtering algorithm based on item recommendation for enterprise users interested in industry application products,design the compensation recommended strategy to solve the problem of the cold start,and testing function and performance of the system in order to check the system's reliability and stability. |
PDF Full Text Request