A Personalized Learning Recommendation Method Based on Deep Reinforcement Learning

Abstract

The invention discloses a personalized learning recommendation method based on deep reinforcement learning. The method comprises the steps of: defining knowledge points and difficulty attributes of questions, and constructing a knowledge point network graph according to relationships among the knowledge points; determining relationships among the questions under the knowledge points according to the relationships among the knowledge points, and constructing a question network graph; obtaining a sub-graph for a current state of a specified user in the question network graph according to user behavior data to use the same as a learning boundary; and then using a deep reinforcement learning algorithm and user history records for modeling, and obtaining, by training, a strategy of how to select a cut set in the sub-graph under the current state of the user. According to the method of the invention, a best question can be intelligently recommended for the user, user learning time can be saved, learning efficiency thereof can be enabled to be improved, and learning experience can be improved.

Description

technical field [0001] The invention relates to the field of personalized learning recommendation research, in particular to a personalized learning recommendation method based on deep reinforcement learning. Background technique [0002] With the launch of more and more Internet education platforms, online learning resources have also been greatly enriched. Users can learn anytime and anywhere, and at the same time get tests at any time. This kind of experience is self-evident for users. . However, the differences in individual differences, interests, and learning styles of students greatly affect the learning effect. There is a low learning efficiency in non-differentiated teaching, and it is difficult to teach students in accordance with their aptitude. American psychologist Noel Tichy (Noel Tichy) once proposed that the most ideal state for a person to learn is to be in the "stretch zone" where the things to be learned are appropriately challenging. Then, mining the us...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art that cannot perform personalized recommendation, and provide a personalized learning recommendation method based on deep reinforcement learning, which can intelligently recommend "learning area" topics for users and save users' learning time , to improve learning efficiency and learning experience

Images