Study On Learning Evaluation Of MOOC Forum Based On Text Mining

The popularization and rapid development of educational informationization have motivated MOOC to become an indispensable component of supplementary education.It is also increasingly valued by educators and researchers.The MOOC discussion area has accumulated a large amount of text data which are a direct expression of learner motivation,effectiveness,and quality.At the present stage,the study of MOOC learning evaluation focuses on the formulation of evaluation indicators and the development of evaluation systems.There are few studies on the evaluation of learners' cognitive level and learning ability while were based on manual evaluation,resulting in a large amount of text data that cannot be fully utilized.So online educators cannot timely and comprehensively understand online learners' learning situations.The emergence of text mining technology has accelerated the processing speed of big data,making it possible to use a large amount of data in the MOOC discussion area.Therefore,it is very urgent to use text mining technology to study the evaluation of MOOC discussion areas.This dissertation uses the posts published by learners in MOOC discussion area as research objects,combines literature research methods,action research methods,etc.to establish a learning evaluation framework in MOOC discussion area.The MOOC Discussion Area Evaluation Framework is used to evaluate the discussion posts,and then the text mining technology is used to automatically evaluate 2770 discussion posts in the MOOC discussion area so as to obtain the cognitive level of the learner's answer.Thus,teachers can evaluate MOOC learners' learning effects,improve learning quality,and reduce the laborious process of manual evaluation.The specific research work of this dissertation is as follows:Firstly,we reviewed related literature such as learning assessment and text mining,summarized their research status at home and abroad in the MOOC discussion area,and introduced the theoretical basis of the dissertation such as Bloom's cognitive target classification theory and text mining technology.Based on Brommer's cognitive target classification theory,the MOOC discussion area learning evaluation framework was designed and the MOOC discussion posts were manually evaluated in this framework.The evaluation framework was designed to establish the criteria for the implementation of the following automatic evaluation.Secondly,the text mining technology was introduced and the overall framework of the automatic evaluation model was established.It was applied to the learning evaluation in MOOC discussion area,and the key technologies of text mining applied in learning evaluation in MOOC discussion area are analyzed which included text preprocessing and classifier construction.Thirdly,the established learning evaluation model in MOOC discussion area was applied to specific instances.Under the Python environment,a text mining-based MOOC discussion area learning evaluation model was constructed.The collected text data which was classified using the evaluation framework proceed with text preprocessing and experimental simulation.Then the evaluation results were obtained.Finally the test set was used to verify the learning evaluation model in MOOC discussion area.Through text mining technology,2770 discussion posts in the MOOC discussion area were simulated.The experimental results show that the model's total prediction accuracy reached 93%,the recall rate reached 93.1%,and the f1-score was 93.4%.The conclusion was that text mining technology can be used to evaluate MOOC discussion area learning.