Modeling and Analysis of Digital Education Resources Sharing Based on Improved Collaborative Filtering Algorithm

With the rapid development of information technology, education informatization has been paid more and more attention by people. In the development process of educational informatization, realizing individualized learning is the urgent need of learners. Personalized recommendation technology can help personalized learning and significantly improve the learning efficiency of learners. Education is the main means of cultivating talents in our country, but the traditional teaching mode is limited by time and place. Therefore, in today's era, teaching through the Internet is convenient and unrestricted. Collaborative filtering algorithm can recommend personalized digital education resources for users according to their hobbies and learning behaviors, so as to improve the utilization rate of resources. Two algorithms, the algorithm of pre-computing user similarity and the algorithm of calculating similarity matrix by introducing time scoring weight, are improved. Experiments show that the algorithm of pre-computing user similarity shortens the time of pushing related information to users, thus effectively improving the calculation speed. Time scoring weight is introduced to calculate similarity matrix to improve the quality of recommendation. After the two algorithms are applied at the same time, the recommendation system has a significant improvement in calculation speed, accuracy and novelty.