Research On Sentiment Tendency Of Online Education Courses’ Comment Text

Online education platform is a new thing,it makes teaching to carry out on the Internet,it attracted a large number of teachers and students carry out teaching activities on this platform because of its learning convenience.With the increasing number of users,online education platforms accumulated more and more the educational data,students’ s activity data recorded all the time,the data had become to a large education sector data,research scholars can predict the performance of students or evaluation teaching by these data.At present,there are a lot of online learning platforms in our country.The construction of the MOOC platform of the major universities are also being carried out in full swing,but the teaching quality of some courses on the platform have not been regulated and perfected.Many students encounter problems when they finishing the course then found not suitable for their owns.So the paper research and analysis the courses comment data of online education platform,it can reflect the students emotional attitude for the completed course,these comments data Emotional Tendency Studies can evaluate the quality of teaching and provide decision-making recommendations for students who wish to choose this course.This research chooses the course comment text under the English module in the Cloud classroom of Netease as the initial data of the experiment,these datas can be crawled by crawler program based on WebDriver.After crawling the data,I process the data such as exclude irrigation,advertising text and so on,Finally,choosed 5000 comments text as experiment required data.After obtaining the experimental data set,this paper carry on the emotion tendency analysis to the course comment text based on the emotional dictionary and machine learning two kinds of text emotion analysis method,the emotion dictionary experiment uses the PMI algorithm,the machine learning experiment uses the SVM algorithm.In the experiment of PMI algorithm,the text is based on the NTUSD emotional dictionary,used default reference vocabulary and the three groups of reference words executed text classification experiments,The experimental results show that the the latter is better.In the experiment of SVM algorithm uses the LibSVM tool to train the training set and test the testing set,used the polynomial kernel function,then the classification effect of the experiment is better than the experiment of default reference word,slightly worse than the experiment of three groups of reference words.After analyzing the experiments,this paper puts forward the scheme to improve the experiment according to its advantages and disadvantages: adding the stop word processing,adding the negative word analysis and SVM kernel function tuning.In this paper,I executed PMI experiment when the negative word become effective,others used SVM experiment,finally,integrated the classification results.The experimental results show that the text classification by this method is better than all of the previous experiment.In this paper,the study of the emotional tendencies of the online course comment texts is emotional polarity classification,not involved for the problem of multi-classification.With the development of large educational data,the study of online course comment texts emotional tendencies will be more and more,which will further enrich the theory of emotional analysis.