Short text content classification method and system

Abstract

The invention discloses a method for classifying short text content, which includes: obtaining short text content of a social network platform; obtaining contextual emotional feature values ​​and prior emotional feature values ​​of the short text content; using model training to generate the short text content Word vector; use multi-window convolution operation to obtain the semantic relationship of the short text content at different granularities, and combine the pooling operation to abstract the semantic representation of the short text content from different levels; use a bidirectional long short-term memory network to obtain Semantic representation of the short text content; and combining emotional feature vectors of different levels to obtain an output vector, using a function to calculate the output vector to obtain a probability value that the short text content belongs to one or more content categories, and using the probability The content category with the largest value is used as the category of the short text content. The invention also discloses a short text content classification system, which can realize the aforementioned short text content classification method.

Description

technical field [0001] The present invention relates to the technical field of information processing, and in particular to a short text content classification method and system based on a deep neural network. Background technique [0002] Emerging social media represented by Weibo has become an important medium for Internet users to obtain news information, social interaction, self-expression, share opinions, disseminate information and social participation. The main platform for dissemination. As of September 2017, the global microblog service site Twitter has 330 million monthly active users, while the Chinese microblogging platform Sina Weibo has 376 million monthly active users and 165 million daily active users. Hundreds of millions of active users come from different social and cultural backgrounds and spread all over the world, generating a large amount of text information containing users' opinions and emotions every moment. [0003] Potential users will understan...

AI Technical Summary

Problems solved by technology

[0004] However, due to the short text of Weibo, incomplete grammatical structure, random and noisy expressions, the classification of Weibo texts faces great challenges.

The existing classification methods mainly use manual methods to construct classification features, and the bag-of-words model is often used for classification features, which is difficult to describe word order information, and faces the problems of dimension explosion and data sparseness

In addition, the text of Weibo often has "explicit praise and implicit criticism" or implicit expressions, and it is difficult to discover and describe such implicit semantic relationships by artificially constructing features.

Images