Research And Application Of Online Learning Community Learners' Emotion Recognition Based On Transfer Learning

With the development of educational informationization,online learning community has become an important way for learners to acquire knowledge in non classroom environment.Online learning community has the advantage of allowing learners to study online anytime and anywhere,which makes up for the unbalanced distribution of educational resources,so that learners can get more learning opportunities and benefit a lot.Because the emotional state produced by learners in the learning process will have a certain impact on their learning effect positive emotion will help to enhance learners'learning motivation and promote the effectiveness of cognitive process,while negative emotion will affect learners' patience and attention and hinder the cognitive process,these emotional factors will eventually affect learners' learning effect,but due to online Under the learning environment,teachers and students are separated in time and space,so it is difficult for teachers to grasp the emotional state of students in the learning process,and there is the so-called lack of emotion.In the context of the lack of emotion in the online learning community,it is worth studying to recognize the emotion of the learners in the learning process,so as to realize data-driven learning diagnosis and learning intervention,and provide personalized guidance for learners.Therefore,the main research work of this paper is as follows.First,construct the emotion recognition model of learners.First,the semantic map of text is constructed by combining the word order information and dependency syntax information of text,and the word vector containing semantic information is calculated by graph embedding technology;then,the emotional features of text are extracted based on transfer learning technology,that is,the shared emotional features of different fields are obtained by using a hierarchical attention network first,and then the shared emotional features of different fields are obtained by using another hierarchical attention network The unique emotional features in the same field are mapped to the shared emotional feature space.Finally,the online learning community learners' comments are identified by combining the two hierarchical attention networks.Second,design and implement the learning analysis system,apply the learners'emotion recognition model to the learning analysis system,as one of the supporting applications of the system.Based on the data in the online learning community,the learning analysis system models and analyzes the behavior,emotion and knowledge of learners,and then provides various services for teachers,students and online learning system,optimizes the teaching process,and promotes the effective learning of learners.In the design,based on the data angle and object interaction angle,from the angle of data mining,this paper constructs a learning analysis system model with students,teachers and online learning community as the core.In implementation,the above-mentioned learning analysis system is implemented based on B/S architecture coding,and different learning analysis applications are encapsulated into independent service modules,so as to improve the overall scalability of the system.In this learning system,as a supporting function module,the learners' emotion recognition model can complete the learning analysis work such as the visualization of the results of emotion recognition of the learners' curriculum review data,the calculation of the polarity distribution of the emotion of the curriculum review,and the recommendation of the potential interest courses of the learners in the online learning community.A comparative experiment is carried out on the real data set of five online courses.The results show that this method can effectively identify the emotion of the learning community course review text,and verify the validity of the emotion recognition model used in this paper.Based on the implementation of the algorithm,the emotion recognition model is further applied to the learning analysis system,as a part of the learning analysis system,to provide data support for Personalized Learning Resource Recommendation and other learning analysis work.