Design And Implementation Of Big Data-based Educational Resources Personalized Recommendation System

As the big data era is coming and personalized education concept becomes more and more popular, how to make teachers and students take advantage of the valuable resources in PB class of electronic education resources becomes one of the most popular issues. Personalized recommendation system is very important in resolving information overhead. It helps people quickly and automatically finding out interesting and valuable information, and is recently used a lot in E-Commerce. However it is still in the exploratory stage in Education without having a mature application system. Thus, this thesis is aim to perform research and implementation on personalized education resource recommendation system which processes big data.The main achievements of this thesis include:1)Based on the functional requirements and non-functional requirements of the personalized education resource recommendation system, B/S and layered structure are designed, the whole design, logical structure and key techniques are planned, and core modules of the system are implemented.2) Based on Hadoop distributed file system and MongoDB technique, secure and efficient big data storage are implemented. Combining with Spark Stream technique, implementation of data module real-time update is done to improve the efficiency of personalized recommendation.3) The thesis built offline, nearline and online recommendation structure by taking advantage of Mahout and MLLib. Flexible big data processing is implemented by using Spark and Hadoop, and better organization resource model and user model are formed by building course classification tree. Also the cold starting problem is solved and the accuracy of the arithmetic is improved by using mixed recommendation arithmetic based on content and cooperation filtering.4) By using Java techniques cross platform server side is implemented, and also interactive user interface is developed by using Spring MVC and Ajax.The education resource personalized recommendation system that this thesis implemented has been used in a university network center, and has met the expectation. The system provides the ability to process big data as it increases, and the recommendation functionality based on the security, high efficiency and real-time of big data. It really helps the students and teachers a lot in terms of taking advantage of valuable education resources.