Text classification method and device

Abstract

The present application provides a text classification method and device, which relate to natural language processing technology and deep learning technology in the field of artificial intelligence. The method includes: acquiring text to be classified and a preset text library; performing word segmentation and part-of-speech tagging processing on the text to be classified to obtain Multiple word segmentation results and part-of-speech tagging results corresponding to each word-segmentation result; match each word-segmentation result with a preset text library, and determine the target word segmentation result and the first weight information corresponding to the target word segmentation result based on the matching result and the part-of-speech tagging result; Based on the first weight information, determine the second weight information corresponding to other word segmentation results in the multiple word segmentation results except the target word segmentation result; according to the first weight information and the second weight information, obtain the text feature information of the text to be classified; The text classification model performs correlation identification processing on the text feature information, and obtains the text classification result of the text to be classified. The present application can improve the accuracy and efficiency of text classification.

Description

technical field [0001] The present application belongs to the field of computer technology, and in particular relates to a text classification method and device. Background technique [0002] Text (eg, official document) classification is a fundamental task in natural language processing. In related technologies, text classification methods based on templates, text topic classification methods based on information retrieval, and text similarity-based classification methods are usually used to classify texts. [0003] The template-based text classification method relies heavily on rules and templates, consumes more system resources in the text classification process, and the constructed rules have low generalization ability and insufficient versatility, resulting in low text classification accuracy. The text topic classification method based on information retrieval retrieves and classifies by training the classification model, but text classification is a multi-classificati...

AI Technical Summary

Problems solved by technology

[0003] The template-based text classification method relies heavily on rules and templates, consumes more system resources in the text classification process, and the built rules have low generalization ability and insufficient versatility, resulting in low text classification accuracy.

The text topic classification method based on information retrieval retrieves and classifies by training the classification model, but text classification is a multi-classification problem, and the imbalance of multi-class data makes the prediction accuracy of the classification model low.

The text similarity-based classification method uses the result of sentence similarity calculation directly on the text as the basis for text classification. However, this method ignores the similarity between the contents of the text itself, and does not distinguish the key information of the text. , resulting in lower text classification accuracy

Images