Research On Sentiment Analysis From The Perspective Of Barrag

Today,the principal contradiction in our society has been transformed into the contradiction between imbalanced and inadequate development and people’s evergrowing needs for a better life.This principal contradiction is also reflected in the field of variety show creation.In recent years,cultural variety shows have been very popular.CCTV has been committed to exploring cultural topics and seeking innovative forms of artistic expression.CCTV has previously produced cultural programs such as“Chinese Character Dictation Contest”,“Chinese Idiom Congress”,“Chinese Poetry Competition”,“Readers” and “National Treasure”,which have rich connotations and profound connotations,and carry forward mainstream values and show cultural confidence.In order to directly implement the basic requirements of cultural construction in the new era in the report of the 19 th National Congress and produce cultural programs different from the common forms on the market,CCTV takes poetry as the breakthrough point,adapts classical poetry into modern pop music,cleverly integrates “poetry” and “song”,and puts the poetry from “memorizing” to “singing out”.It launched a new cultural music program called “Everlasting Classics”,calling on the people to pay attention to and inherit traditional Chinese culture.The show has sparked heated debate since its release,scoring 9.3 on Douban the day after its debut.With the continuous development of the Internet and media,bullet screen,as an emerging form of comment,is now the mainstream mode of commenting on video websites,which can timely reflect the emotional attitude of the audience.First of all,this paper obtains the bullet screen data of the third and fourth seasons of “Everlasting Classics” on the i Qiyi website through python.After data cleaning,Chinese word segmentation and other preprocessing,a word cloud map is drawn by counting word frequency and TF-IDF value,and the general situation of the bullet screen in the two seasons is preliminary obtained after the analysis by the singer.Secondly,the bullet screen was manually coded and the data set was divided.The prediction and comparison were made according to the three machine learning models of support vector machine,naive Bayes,random forest,and BERT deep learning model,and the BERT model was the optimal model.The experimental results show that the accuracy rate,recall rate,F1 value,accuracy rate and AUC value under ROC curve of BERT model in the third quarter are 0.8616,0.8754,0.8672,0.8046,and 0.83.The accuracy rate,recall rate,F1 value,accuracy rate,and AUC value of the BERT model in the fourth quarter were 0.8544,0.8379,0.8451,0.7865,and 0.70.Thirdly,with the help of the LDA model,theme mining under emotion classification is completed to clarify the positive and negative themes of the program.According to the results,most of the audiences have a positive attitude toward the program,but some of them have negative emotions.Most people focus on the superficial aspects of stars and music.Even though the show’s situation improved slightly from the third season to the fourth season,the focus gradually shifted from the surface to the inside,but the real focus on the content of the show--poetry and traditional Chinese culture itself is rare.Although the program has obtained a high audience rating and great influence,the program producer should pay more attention to cultural inheritance and the establishment of cultural confidence while paying attention to audience rating and flow rate.Finally,based on the results of emotion analysis,the paper draws a conclusion and puts forward some suggestions for the development of the program to promote and spread Chinese traditional culture.