Design And Implementation Of Open Platform For Personalized Recommendation

With the rapid development of Internet technology, which lets the people entered the era of big data. In mobile internet today, people do not just browse data, but also create data. In the face of the vast amounts of information, it’s hard for them to find the resource which they really like. Therefore, the potential value of mining data and the personalized service for the users have become a hot topic of big data era. Meanwhile, personalized recommendation technology arises at the historic moment. Personalized recommendation is based on the user’s purchase behavior, and the characteristics of the interest, recommend information of interest to the user. The essence of the recommendation technology is based on the analysis of the user’s behavior data to predict the future behavior of the user’s preferences, and to predict the results of the show in the form of a valid user.Open platform is an internet platform architecture idea in recent years. Such as Open Platform, which is the earliest domestic large-scale used. The server-side open platform, a kind of computing power encapsulated into a set of computer is easy to identify data interface, used for third party users. Through the open platform, the caller can implement complex interface data interaction capabilities, so as to develop the rich applications.This thesis combines the personalized recommendation technology and the ideasof open platform architecture, the personalized recommendation and open platform architecture was proposed. The personalized recommendation services are offered in the form of interface by the platfrom, mainly including the recommended interface and configuration interface. The caller of the interface do not need to purchase the servers and program to implement recommendation service, and only need to use the interface to access the recommendation service, by this way, it greatly reduces the cost of development. At the same time, any caller has the authority to the interface can access to the service in anywhere which has the internet. The callers just need to pass the user id and can get the personalized recommendation items through the recommendation interface and can also start the service through the configuration interface.In terms of personalized recommendation algorithm, it paper presents a kind of loose coupling algorithm programming idea. With this idea, reduce the coupling algorithm and platform greatly, update algorithms library efficiently and integrated more algorithm.In open platform technology architecture aspect, it follows the interface-oriented design principles. Restlet is used as interface layer, Kafka is used as communication middleware layer between platforms, Redis is used as database layer to provice interface service, and Hadoop is used to implement distributed computing and storage. The collaboration layers, it completes the function of open platform and achieves the personalized recommendation of open platform.