Student learning preference identification method based on multi-modal fusion

Abstract

The invention discloses a student learning preference identification method based on multi-modal fusion, and belongs to the technical field of teaching, and the method comprises the following steps: S1, obtaining the voice information of a student through a sound sensor, and carrying out the voice feature learning of the obtained voice information through a voice feature learning network, S2, obtaining the vital sign information of the student through a wearable device, and carrying out dimensional vital sign learning on the vital sign information by adopting a dimensional vital sign learning network. According to the invention, a multi-scale mode is adopted, n small videos with different time lengths are segmented according to video background changes, accurate recognition is realized, multi-modal fusion is adopted, multi-modal features are established according to expression, voice, heart rate, breath, blood pressure and other information, the recognition effect is better, network combination is realized at a deeper mechanism level through the LSTM model, and the long-term memory ability is enhanced.

Description

technical field [0001] The invention relates to the field of teaching technology, in particular to a method for identifying students' learning preferences based on multimodal fusion. Background technique [0002] The education of students in primary and secondary schools is the education stage that can best reflect and even determine the direction of students' interests. At present, parents and teachers of students attach great importance to cultivating students' interest, and select learning resources from a huge library of learning resources to help cultivate students' interest in learning. This method will not only cause waste and unreasonable occupation of learning resources, but also self-determination. The recommended learning resources are numerous and complex, and it is difficult to grasp the interests of students in learning, so as to achieve the effect of enhancing learning interest; in addition, the educational resources provided by the school cannot be personaliz...

AI Technical Summary

Problems solved by technology

At present, parents and teachers of students attach great importance to cultivating students' interest, and selecting learning resources from a huge learning resource library to help cultivate students' learning interest will not only cause waste and unreasonable occupation of learning resources, but also independently The recommended learning resources are numerous and miscellaneous, and it is difficult to grasp the learning interests of students, so that it is difficult to achieve the effect of improving learning interest; in addition, the educational resources provided by the school cannot be personalized according to the abilities and interests of student users, which leads to The utilization rate of existing educational resources is generally not high. Even if the content in the educational resource library is richer, it will have little impact on the learning of primary and middle school students.

[0003] The traditional teaching design does not consider the differences of students, but the actual situation is that there are huge groups of students, and there are differences in learning preferences. It is difficult for the model to take into account the individual needs of all students

Images