Exploration of Innovative Teaching Methods in the Teaching Reform of Linear Algebra Course in the Internet Era

Abstract

Abstract For the task of ternary extraction in unstructured text, an entity recognition model based on BiLSTM-CRF is proposed. By introducing graph convolutional neural network and attention mechanism respectively, a relation extraction model based on attention-guided graph convolutional neural network and a text classification model based on GCN are proposed to realize the automatic extraction of linear algebra knowledge triples. To solve the problem of ignoring semantic information of knowledge points and sparse data, the linear algebra knowledge level matrix of learners is obtained by using the DKVMN model. The knowledge graph’s relationships are assigned weights by experts through evaluation. The semantic information in the knowledge graph and the weights of relationships between nodes are utilized to complement the knowledge level matrix as a way to recommend appropriate linear algebra exercises for learners. Compared with the traditional teaching class, the experimental class of this paper’s method scores higher in algorithmic thinking, collaboration, critical thinking, and problem-solving by 0.67, 0.27, 0.41, and 0.37, respectively. Based on the learning data of each chapter’s knowledge points, it is understood that the standard deviation of the eigenvalues and eigenvectors in the eigenvalue and eigenvector and quadratic form, quadratic form, similarity matrices, positive qualitative and geometric questions, and proofs are respectively 4.076, 3.226, 6.185, 3.292, and 4.105, and the standard deviations are generally large compared to other chapters. This paper demonstrates that the method can be taught accurately depending on the students’ linear algebra level.