Online education-oriented learner abnormal learning state prediction method

Abstract

The invention discloses an online education-oriented learner abnormal learning state prediction method. The method comprises the steps of preprocessing high-dimensional online education platform log information and learner registration information, and coding and constructing learner portrait features based on a self-supervised learning method; constructing state features of the learner, further constructing a state feature sequence based on a generation time sequence of the state features, and constructing a state feature graph based on cosine similarity among the state features; constructing a long-short term memory-graph attention deep network conforming to learning badness degree prediction of online education, and determining the number of layers of the network, the number of neurons of each layer and the dimensions of input and output; constructing a pseudo tag based on the noise tag to perform iterative training on the network; and predicting the abnormal learning state and degree of the learner in the to-be-predicted learning stage by using the trained network. According to the method, the state abnormity degree of the learner is predicted by using the learner registration information and the learner log information, and a reference is provided for a teacher to carry out targeted guidance and help on the learner.

Description

technical field [0001] The invention relates to the technical field of online education, in particular to an online education-oriented method for predicting abnormal learning states of learners. Background technique [0002] With the widespread popularity of modern computer networks and the rapid development of home electronic equipment terminals, online education using modern technologies such as computer networks and artificial intelligence has become an important part of family education. In 2020, the market size of the online education industry will increase by 35.5% year-on-year. However, as a means of distance education, the characteristics of online education with less interaction between teachers and students and difficult learning supervision have caused a negative impact on the teaching effect that cannot be ignored. In recent years, the number of online education participants and online courses have increased year by year, and a large amount of online education le...

AI Technical Summary

Problems solved by technology

In 2020, the market size of the online education industry will increase by 35.5% year-on-year. However, as a means of distance education, online education has less interaction between teachers and students and difficult learning supervision, which has a negative impact on teaching effects that cannot be ignored

[0008] The above literature methods are used to find learners with abnormal learning states. The main problems are as follows: First, both literature 1 and literature 2 are aimed at finding problems for poor learners. Individualized instruction is particularly important, for example, learners with occasional and recurring learning misbehaviors require varying degrees of supervision and instructional methods; secondly, learning outcomes or test scores for a semester or course do not adequately reflect learners' progress in the learning process. Whether there is a need for targeted supervision in the course, the test scores depend on the different emphases of the course exam questions, and students who study more seriously in the part of the exam questions will get higher scores, which leads to the academic performance not fully reflecting the current teaching part of the learners. Learning status; in online education, teachers hope that learners with abnormal learning status in any teaching part can be discovered in time and provide targeted guidance, so directly using learning grades as training labels to measure the learning status of the whole process introduces label noise; At the same time, Literature 1 and Literature 2 mine learner portrait features or temporal representations for learning. The lack of attention to different perspectives of features also makes it difficult to carry out label disambiguation based on multi-view features to solve the noise problem.

However, obtaining accurate labels of bad learning status and its degree requires a lot of manual labeling costs. How to construct labels based on existing academic performance and multi-view features to more accurately evaluate the abnormal learning status and degree of learners at each stage has become a problem. a burning problem

Images