Research On Teaching Comment Sentiment Analysis Based On Association Mining

With the rise of the popularity of the Internet,various online learning methods such as online live teaching,micro-classes and MOOC have mushroomed.In the field of education,teaching evaluation is an extremely important part of teaching activities.It is mainly to make value judgment on teachers' teaching process and students' learning effect,so as to make teaching improvement targeted.At present,all kinds of network learning technology constantly improving,but the teaching evaluation is relatively lag,cause each large network learning platform,although has the potential of schools of thought contend,content has ramifications,cannot let students quick positioning,in order to improve the quality of the online classroom,improve the students' interest in learning,strengthen the link of teaching evaluation is urgent.At present,the teaching evaluation of online learning platforms is mainly in the form of rating and comments.For a large number of comments,it is an inevitable trend to use machine learning method to conduct sentiment analysis in order to quickly locate the evaluation content.Comprehensive research status at home and abroad,the present research mainly adopts the analysis of the sentiment SVM or ANN,such as machine them,no matter which method is adopted to improve the sentiment analysis research,research method itself has certain defects,which affect the operation process,or cannot get ideal results.For example,support vector machine(SVM)algorithm is difficult to simulate training for large-scale samples,and performs poorly in multi-classification problems.The range of topics covered in this paper is mainly online course learning.This study aims to explore new methods of emotion analysis and expand the diversity of emotion analysis methods,so that researchers can have more choices when facing different types of research samples and choose research methods more suitable for the research samples.The method of this study is to determine the best classification of open feedback in the context of teaching emotion by obtaining feedback comments of open online teaching and adopting a new emotion classification method,and classifying emotions in context into positive or negative ones.This research adopts a new method based on association mining called affective phrase pattern matching.The object of the model analysis is the open comments on the courses of the online learning platform,from which participants' comments on the courses are extracted,and students are allowed to feedback the factors affecting teaching and learning to teachers or the platform.The methodology adopted consists of four main phases:(1)collect feedback data through the sentiment dictionary and perform marking;(2)Sentiment phrase pattern matching(SPPM)was used to analyze sentiment phrases based on association mining method.In combination with frequency of sentiment phrases,twoway forward traversal was used to separate multiple phrases in the teaching feedback sentences.(3)conduct sentiment analysis based on the affective polarity score of teaching affective dictionary.(4)support vector machine(SVM)algorithm was used for secondary analysis of the data,and then horizontal comparison was made with the experimental results of this study to verify the results of sentiment analysis and verify the effectiveness of the experiment.The advantage of this study mainly lies in the maneuverability of the method.Most of the previous studies were too specialized to be popularized.The experimental results of this study are composed of four evaluation indexes(accuracy,precision,recall and F1 score).The results obtained by the method adopted in this study are superior to the performance of support vector machine(SVM)algorithm in many aspects.The comparative experiment proves that the method adopted in this study,pattern matching of affective phrases,can achieve the real purpose of affective analysis in the aspect of text affective analysis.In the research process,it is found that there is a big gap between domestic and foreign emotion analysis research,mainly in the aspect of emotion dictionary,the classification is not detailed and diverse enough,the lack of polarity value evaluation,the word bank is not updated enough.It is hoped that future researchers can strengthen the research in this field and promote the sound development of text emotion analysis.