Research And Application Of Personalized Recommendation Of Teaching Resources Based On Collaborative Filtering

Internet has become an important way for people to reach information. It provides people with abundant information resources. As the surge of information quantity, search technique is facing the big challenge. It becomes difficult for people to find the right and suitable resources, so the traditional search technique is hard to meet the personalized requirements of users. Under this circumstance the recommend technology arises and shows strong vitality. Recommend technology make recommendation that interest users based on users'information, it can help people find the information they need conveniently.This paper applies recommend technology into educational informatization field. Teaching resource plays an important role in teaching. There are lots of teaching resources in network teaching platform. Usually, the platform provides the same user interface to every single user. But users have different needs because they have different interests.Collaborative filtering is the most widely used among all the recommend technologies. It helps users discover potential interests and then make recommendations. Users will reach the resources accurately and rapidly among large quantity of teaching resources. The personalized needs of users are being met.This paper researches of users from an online teaching platform, analyzes the feature of this specific group, designs a recommendation procedure for them, and improves the similarity computation to have a better recommend quality. To avoid sparse matrix problem in collaborative filtering, a hybrid recommendation algorithm has been designed, that is combined with user-based and item-based recommendation. By constructing a recommendation engine based on Apache Mahout, the improved algorithms are being implemented. The test on the algorithms proved the effectiveness of improved and hybrid recommendation algorithms.