Microblog sentiment analysis method based on expression dictionary and sentiment common sense

Abstract

The invention provides a microblog sentiment analysis method based on an expression dictionary and sentiment common sense. The method comprises the steps that microblog text data under a certain topicis collected, preprocessed and subjected to word segmentation operation; selecting high-frequency use expressions to construct a microblog expression dictionary; binary matching of the ConceptNet semantic library is extracted, emotion labeling is conducted, and extension is conducted through a synonym dictionary to form an emotion common sense library; and performing sentiment analysis on the microblog according to the emoticons and the weight calculation of the sentiment common sense. According to the method, the emotion classification of the microblog is judged in a mode of combining the expression dictionary and the emotion common sense; explicit features such as network words and emoticons and implicit features such as sentiment common senses are fused in the sentiment analysis task,and implicit sentiments to be expressed by microblog texts can be deeply mined to a great extent, so that the accuracy of sentiment analysis is improved.

Description

technical field [0001] The invention belongs to the technical field of text emotion analysis in natural language processing, in particular to a microblog emotion analysis method based on expression dictionaries and emotion common sense. Background technique [0002] Weibo has gradually become an important emerging social network platform in the Internet age. Users can share personal life, publish personal views, and interact with friends on Weibo through web pages or client terminals. So far, the number of microblog users has exceeded 300 million. In the massive microblog data, there are many microblog resources with subjective emotional tendencies of users. Research on how to efficiently mine these complicated microblog messages The themes and sentiments in the content are helpful to the government's public opinion analysis and network supervision, and also help enterprises and institutions to guide public opinion on topics they care about. [0003] However, due to the uni...

AI Technical Summary

Problems solved by technology

[0003] However, due to the unique originality and unpredictability of Weibo, the existing sentiment analysis of Weibo basically calculates the emotional value from the explicit features such as emoticons and network words. The judgment of emotional tendency has an important influence, and the expression of implicit emotion does not contain emotional words to a large extent. It requires readers to have a certain knowledge background and use reasoning to discover some hidden emotions implicitly conveyed by common sense.

Images