Micro-blog online emergency detection method based on emotionl analysis and tagging

Abstract

The present invention belongs to the field of network detection, and specifically to a micro-blog online emergency detection method based on emotion analysis and tagging. The method comprises: by using an emotion classification model emotion wheel, constructing an emotion analysis model - emotion co-occurrence map; by using the emotion analysis model constructed in step (1), performing emotion classification on micro-blogs in a micro blog flow, and detecting an emergency period of the micro-blog flow by using a kleinberg algorithm; extracting micro-blog tags during the emergency period, filtering out a junk tag, and performing word segmentation on remaining tags; forming an initial keyword of an event; and by using the keyword generated in step (3), extracting a word related to the keyword in a micro-blog, to form a final description of the event. According to the method provided by the present invention, the emotion co-occurrence map based on the emotion wheel is constructed, so that the emotion classification is more detailed, emotion is easier to understand and explain, and compared with events based on emotion symbols, detection accuracy is higher.

Description

technical field [0001] The invention belongs to the field of network detection, and in particular relates to a detection method for microblog online emergencies based on sentiment analysis and labels. Background technique [0002] In recent years, with the vigorous development of Web2.0 technology, a series of social networks have emerged. These social networks such as Sina Weibo and Twitter attract a large number of users. Users are active on social networks, posting a large number of Weibo messages, which contain views or opinions on certain events. By mining these microblog messages, a large amount of deeper information such as user emotions can be obtained. The use of these in-depth information can provide services for the government or enterprises. For example, the government can use this information to judge whether people support laws and bills, and what kind of views they hold on a certain social event, so as to control and guide public opinion; enterprises can By...

AI Technical Summary

Problems solved by technology

However, these two methods are not applicable in the case of short microblog text

First of all, the amount of microblog data is large, and it takes a lot of time to extract feature words and form a tfidf matrix for each microblog

Secondly, microblog expressions are irregular and changeable, and may contain a large number of new words, forming a sparse matrix, which is not conducive to the calculation of similarity and increases the difficulty of recognition

At the same time, the traditional method only completes the extraction of emergencies, but does not conduct deeper analysis of emergencies, such as sentiment analysis

Images