The Design And Implementation Of Recommendation System Of Teaching Resources Based On The Learning Style Under The Cloud Computing Environment

The 21 st century is the era of Internet. "Internet + education" can make the students’ learning style changed. The use of the Internet transfers learning methods from physical classroom to classroom network, from face-to- face answering of the teacher to searching online information to solve problems. Students not only centralize in listening to the teacher in class,but also learn diversify on the Internet so that they could arrange their own study plan reasonably to improve their efficiency.However, with the rapid development of Internet, there will be explosive growth in the data information.It is significant for us to dig out the teaching resources that students may be interested in or in need in the massive data and then recommend these teaching resources to them according to the characteristics of students themselves in order to improve the student’s learning efficiency and learning effect.Differentstudents will form different learning styles according to their own characteristics. If the learning styles of the students are approximate, their concern of teaching resources will also be similar. Collaborative filtering recommendation algorithm based on item is established on the theory the items of similar crowds pay attention to are similar. This article selects the theoreticallymature and practical collaborative filtering recommendation algorithm based on item to establish the teaching resources recommendation system so that it could recommend support of teaching resources to the students of different learning styles.In this paper,through the following two kinds of methods to improve the performance of collaborative filtering recommendation system and the accuracy of recommendation. First before users of teaching resources are recommended to the students, student users will be grouped on basis of the test results of learning style scale by Felder- Silverman and then be recommended separately for each group in order to enhance the accuracy of recommendation. Simultaneously, with the giant increase of the number of the students’ users and the number of teaching resources, the article used Hadoop parallel computing framework to develop teaching resources recommendation system and adopted a CUR dimension reduction to optimize the collaborative filtering algorithm based on items.With concrete implementation in the form of MapReduce programming model and the operation in distributive computing mode, more tasks nodes quickly finished the computing task. The data matrix eventually merged recessing in the Reduce task node and produced an available list of candidate recommendation so that the recommended system in the article can be suitable for large-scale data analysis computing tasks. Last,according to the accurate rate, recall rate, coverage rate, the collaborative filtering recommendation algorithm and the optimized algorithm are compared in the cloud platform, Through the analysis of the chart, we can see in the case of a large number of users, the optimization algorithm makes the effective performance improvement compared with the before optimization.