Research On Micro Blog Comment Oriented Fine-grained Sentiment Analysis Method

Weibo is one of the most popular social platforms in China.It has the character-istics of fast information transmission and wide coverage.People can comment on a certain event on Weibo to express their emotions and attitudes.Judging the emotional tendencies of users’ comments is not only conducive to the monitoring of management departments,but also has very high application value for rumor suppression、public opinion guidance、and marketing.Weibo comments are mostly presented in the form of text,and it is stipulated that the number of comments should not exceed 140 characters,which is a type of short text.There are the following problems for Weibo comments:First of all,the short length of Weibo comments leads to less effective information that can be captured,resulting in sparse sample features,and it is difficult to accurately extract features in terms of feature extraction;Secondly,Weibo comments are free play,and there is no requirement for grammar,so many comment sentences do not conform to grammatical regulations,which also adds a certain degree of difficulty to text classification;Finally,there are a large number of popular words and Internet terms in Weibo comments,and its update speed is fast,which leads to an increase in text noise.In this paper,sentiment analysis method of microblog comments is researched for the above problems.The main works are as follows:(1)Comment data are crawled on Sina Weibo us-ing crawler technology,and performing data preprocessing,including filtering invalid characters,word segmentation,removing stop words,etc.,and finally a Weibo com-ment corpus resource is formed that can be used for experiments.(2)The text was vectorized using the Bert model,it can fully consider the information on the left and right sides of the word,so as to obtain a deeper vector representation containing up-per and lower semantic information.(3)According to covering complicated vocabulary and unobvious grammatical features of Weibo comment data,this paper proposes a TextCNN-BiLSTM fusion model based on the improved Adaboost algorithm.TextCN-N model can perform local feature extraction,and BiLSTM model can perform global feature extraction,Using TextCNN model and BiLSTM model to process comment data in parallel.Finally,the classification results of the two models are fused through the improved Adaboost algorithm to enhance the classification effect and improve the classification accuracy.At the same time,a multi-level attention mechanism is intro-duced on the basis of the TextCNN model in order to reduce the influence of noise.(4)The classification model with multiple deep learning classification models proposed in this paper was compared and verified,and analyze and summarize the experimental results.