Text classification method, device, medium and apparatus based on convolution neural network

Abstract

The invention provides a text classification method, device, medium and apparatus based on a convolution neural network, wherein, the method comprises the following steps: obtaining a word vector matrix of a text to be classified related to network public opinion; constructing an initial feature matrix according to the word vector matrix, and using the initial feature matrix as the input of the trained text classification model and input the initial feature matrix to the first sequential region block, and determining the output of the region block is determined; The input of each hidden layerin the region block comes from the output of all other hidden layers in the region block. The output of the current region block is taken as the input of the next region block until the output of allthe region blocks is determined, and the output of all the region blocks is transmitted to the full connection layer, and the classification result is determined according to the output of all the region blocks. The network structure adopted in the method can make the transmission of network features and gradients more effective, avoid the problem of gradient disappearance caused by the layer-by-layer transmission of loss function information, and ensure that the network depth can be enlarged while the problem of gradient disappearance can be avoided.

Description

technical field [0001] The present invention relates to the technical field of text classification, in particular to a convolutional neural network-based text classification method, device, medium and equipment. Background technique [0002] With the development of mobile Internet technology, network information is growing explosively, and the network is filled with a large number of useful or useless texts; Different views or Internet opinion. Due to the huge amount of network text information, it is necessary to classify the network information quickly and accurately. As one of the key technologies of natural language processing, text classification can effectively solve problems such as information clutter, and is widely used in tasks such as search engines, spam filtering, personalized news and data sorting. [0003] The current text classification model is generally based on the bag-of-words model and the cyclic neural network model, while the bag-of-words model does ...

AI Technical Summary

Problems solved by technology

[0004] The present invention provides a text classification method, device, medium and equipment based on a convolutional neural network to solve the defect of the gradient disappearance problem in the existing text classification model using the convolutional neural network

Images