A Study On The Automatic Evaluation Of Learners' Online Discussion Quality Based On Semantics

The arrival of the Web2.0 era has brought an opportunity for the development of online education.A large number of online education websites have emerged,and the scale of online learners has increased dramatically.Online discussion forum as an important platform for instructional interaction in online education websites,the quality of online discussion is closely related to the quality of teaching and some scholars have suggested online discussion quality evaluation as part of teaching evaluation.Evaluation of Learner's discussion quality has become a hot issue in educational circles.Although the current research on the evaluation of online discussion quality has paid close attention to the semantics of the discussion content,the way of implementation is still artificial,which has the problems of low efficiency and nonobjective.In this paper,the content analysis method is used as the research method of online discussion quality.Based on the semantic similarity,this paper makes semantic analysis on the content of the learner's post,then calculates the semantic similarity between the content of the learner's post and course content,and weights the semantic similarity of different topics in the course and then calculates the results in order to realize the automatic evaluation of learners' online discussion quality based on semantics.The main work of this paper includes:1.Summarizes the research on the quality evaluation of online discussion.This paper summarizes the relevant studies of evaluation of online discussion,uses the research methods of evaluation discussion quality for reference,clarifies the evaluation angle of this paper,and proposes an automatic evaluation of learners' online discussion quality based on semantics.2.An automatic evaluation system for learners' online discussion quality is designed and implemented.The key technology used is the text semantic similarity algorithm.Firstly,it is necessary to collect the discussion content and related data from the online education website.Secondly,we preprocess the text data,select the feature items which can represent the topic of the text according to the word frequency,part of speech and word length.Finally,calculate the similarity between the two text feature items,and the result is the basis for the online discussion of quality evaluation.3.The consistency between the experimental results and the expert scoring results is verified,and the test results are analyzed and explained briefly.The results show that the experimental results of the semantic-based online quality evaluation system for learners are consistent with those of the expert score.