Abstract: Traditional methods for the sharing of teaching resources often have problems such as unclear resource classification and low sharing efficiency. For this reason, a design scheme of teaching resource sharing platform based on the VEM ( Variational Expectation Maximization) framework is proposed. The platform consists of 4 levels, which are the preprocessing, feature extraction, classification and sharing of teaching resources for collaborative work. In the preprocessing stage, redundant information is deleted to reduce the system storage space. The feature extraction module uses the semantic adjacency matrix to calculate the feature value of the vocabulary in the resource, and selects the vocabulary with a larger feature value. In the classification module, the upper and lower relationship is used to construct a classification framework, calculate the strength of the relationship between each word pair, and realize the classification according to the different strength. The resource number of the completed classification is stored separately, and the teacher can directly find and complete the resource sharing. The experimental results show that the classification accuracy of the teaching resource sharing platform is 98. 6% , and the average time for resource sharing is 1.15 s, proving that the designed teaching resource sharing platform is accurate and efficient.

Key words: variational expectation maximization (VEM) framework; , sharing of teaching resources; , resource preprocessing; , the relationship between superior and subordinate; , sharing efficiency