Text classification method mixing long-short-term memory network with convolutional neural network

Abstract

The invention discloses a text classification method mixing a long-short-term memory network with a convolutional neural network. The advantage of the bidirectional long-short-term memory network in the aspect of learning context information of a text and the advantage of the convolutional neural network in the aspect of learning local features of the text are fully combined; the context information of words is learnt by utilizing the bidirectional long-short-term memory network; the local features of word vectors of the context information are further learnt and extracted through the convolutional neural network; contexts of the local features are learnt by utilizing the bidirectional long-short-term memory network to form a fixed-dimension output; and finally classification output is performed through a multilayer sensor. The accuracy of model classification can be further improved; relatively high universality is achieved; and good effects are achieved in a plurality of tested corpus libraries.

Description

technical field [0001] The invention relates to the field of natural language processing, in particular to a text classification method of a hybrid long-short-term memory network and a convolutional neural network. Background technique [0002] Automatic text classification based on machine learning is the most popular research direction in the field of natural language processing in recent years. and in-depth applications. The so-called automatic text classification refers to the process of automatically determining the text category after analyzing the content of the text by using machine learning under the premise of a given classification system. Before the 1990s, automatic text classification was mainly based on knowledge engineering, that is, manual classification by professionals, and its disadvantages were high cost, time-consuming and labor-intensive. Since the 1990s, many researchers have begun to apply various statistical methods and machine learning methods to ...

AI Technical Summary

Problems solved by technology

Before the 1990s, automatic text classification was mainly based on knowledge engineering, that is, manual classification by professionals, and its disadvantages were high cost, time-consuming and laborious

Existing research and applications have proved that recurrent neural networks are suitable for learning long-term dependencies between language units in sentences, and convolutional neural networks are suitable for learning local features of sentences, but current research has not fully combined recurrent neural networks and Convolutional neural networks have their own advantages, and they do not take into account the context information of language units in sentences

Images