Microblog comment sentiment classification method based on emoticons and text information

Abstract

The invention belongs to the field of microblog comment sentiment classification, and provides a microblog comment sentiment classification method based on emoticons and text information, which comprises the following steps: firstly, preprocessing microblog comments, constructing an emoticon weighting network, calculating five-dimensional sentiment values of the emoticons, and constructing an emoticon sentiment description dictionary; obtaining emoticon vectors in combination with the emotion dictionary, wherein each microblog comment at least comprises one emoticon, and the text length is at least 5; secondly, sending text sentences and expression sequences in the microblog comments into the bidirectional LSTM model, and obtaining text sentence representation and expression graph representation; then, performing learning in combination with text sentence representation and emoticon representation to obtain final embedding of microblog comments, and performing sentiment classification on the microblog comments by using a Softmax classifier. According to the method, the emoticons and the text sentence information can be organically combined, the deep learning bidirectional LSTM model is utilized to carry out microblog comment sentiment classification, and the accuracy of microblog comment sentiment classification is ensured.

Description

technical field [0001] The invention relates to the field of sentiment classification of microblog comments, in particular to a method for sentiment classification of microblog comments based on emoticons and text information. Background technique [0002] Weibo users widely use emoticons and text information to express personal emotions. Emoticons are an important supplement to the emotional expression of text information. The combination of the two can correctly convey the true emotions of Weibo users. [0003] The task of sentiment analysis is to mine users' attitudes and emotional tendencies from various comment data on the Internet. Through the sentiment analysis system, unstructured information can be transformed into structured information. However, emoticons express rich emotions, which are not easy to mine and process. Traditional microblog sentiment analysis generally does not consider the emotional information contained in emoticons, and the accuracy of sentiment...

AI Technical Summary

Problems solved by technology

However, emoticons express rich emotions, which are not easy to mine and process. Traditional microblog sentiment analysis generally does not consider the emotional information contained in emoticons, and the accuracy of sentiment analysis is not high enough.

Images