Multi-dimensional dynamic adaptive exercise recommendation method based on latent semantic model

Abstract

The invention provides a multi-dimensional dynamic self-adaptive exercise recommendation method based on a latent semantic model. The method comprises the following steps: constructing and associating a subject knowledge graph and a subject knowledge question bank; forming the test questions of the subject knowledge question bank into a hierarchical question bank, and selecting basic test questions for testing; analyzing a testing result, and judging an expected reaction mode which the student most possibly has by using posterior probability so as to form a cognitive structure for the student; on the basis of the cognitive structure for the student, adding hidden factor items representing ignored attributes of the student by combining a logic regression algorithm in traditional machine learning and a latent semantic model in a matrix decomposition algorithm, and acquiring a final question prediction model by using a gradient descent method. According to the above technical scheme, hidden feature information is extracted from the answer conditions of the student based on the latent semantic algorithm, and a prediction model of the knowledge point mastering condition of the student can be effectively constructed; and adaptive feedback and personalized recommendation are combined, so the requirements of educating different students in different ways and teaching students in accordance of their aptitude are met.

Description

technical field [0001] The invention relates to a computer application technology, in particular to an intelligent recommendation method, in particular to a multi-dimensional dynamic self-adaptive exercise recommendation method based on a latent semantic model. Background technique [0002] With the gradual maturity of artificial intelligence technologies such as machine learning, image recognition, natural language processing, and speech recognition, the application of artificial intelligence in the field of education has gradually penetrated into the entire process of "teaching, learning, practice, testing, evaluation, and management" education and teaching. , including typical scenarios such as photo search, intelligent assessment, intelligent class scheduling, intelligent correction, and educational robots. As a key technology that can go deep into the core links of education, this kind of self-adaptive learning can independently provide customized help for students with...

AI Technical Summary

Problems solved by technology

Matrix factorization algorithms usually have high accuracy, but hidden factors are not interpretable

Images