Research On Emotional Analysis Based On Weibo Comments

In our daily life,the Internet is becoming more and more popular,and the frequency of people using the Internet is also rising rapidly.For example,people can surf the Internet through web pages or micro-blogs,watch other people's comments,and make corresponding comments on them.Every day,a large number of meaningful comments are generated,from which we can perceive the information of the author,and at the same time,we can know the opinions and thoughts of the author and so on.In China,there is a lot of research on emotion analysis,so some people start to study the comments on weibo.After collecting a large number of literatures related to microblog research,we found that the most commonly used research method is the emotion dictionary,which is to analyze the polarity of microblog comments by constructing corresponding microblog emotion dictionary.Another way is to use machine learning,which is to use some building models to judge the positive and negative of words.As we all know,Chinese culture has a long history,and the language meaning is very rich,much higher than the complexity of English words,many words and the meaning of the word composition is a lot of difference,for there are a lot of related words of the meaning of the word,sometimes using some model algorithm is not very good effect.On the other hand,in the classification process of emotion polarity,the classification method based on emotion dictionary hardly CARES about the vocabulary in the topic field of microblog,thus affecting the correctness of emotion classification results.At present,in China,few literatures have in-depth studies on the emotion analysis of microblog,such as the emotional intensity of the speaker is usually ignored,some algorithms are slightly deficient,and it is difficult to accurately extract the characteristic terms.Collecting the literature of other researchers in weibo comment research and careful check after,found that adopt the method of emotional dictionary is a good choice for sentiment analysis research,and weibo has its own particularity,is a little different with other reviews research association,so we set up exclusive emotional dictionary in weibo,select the relevant weibo comments,improve emotion classification accuracy.After we get the comment text preprocessing,first for it to make it convenient and follow-up studies,then use our exclusive emotional dictionary in weibo,feature extraction and so on to operation,and the corresponding processing negative words,adverbs of degree,weibo emoticons,emotional words and evaluation objects weibo comments.Finally,some algorithm formulas will be used to classify the positive and negative data that have been processed before,so as to achieve an accurate classification effect.We will select some hot topics on weibo.The experimental data include three fields: life,traffic accidents,and science and technology in the field of weibo topics.The final results are reasonable through experimental verification.