Collaborative Filtering Recommendation of Online Learning Resources for E-commerce Logistics Talent Training

The rapid development of online teaching has led to a surge in the number of online learning resources, but students are also prone to the problem of “information loss”. To solve this problem, Collaborative filtering recommendation is an effective method. However, traditional similarity calculation methods take into account the differences in user interests over time, resulting in low recommendation accuracy. Therefore, a recommendation model of online learning resources Collaborative filtering for e-commerce logistics talent training is proposed. Firstly, collect data on users’ online learning behavior and introduce time factors to simulate dynamic changes in user preferences. Then, an improved user resource rating matrix is constructed, the similarity between users is calculated and sorted, and a set of user neighbors is constructed. Finally, predict the target score by setting the target user’s neighbor’s score for a resource and the first N resources with the maximum value, and recommend this resource to the target user, thus completing the Collaborative filtering recommendation of online learning resources. The experimental results indicate that the recommendation model studied has a greater coverage and smaller PMAE in applications, demonstrating stronger recommendation capabilities.