Text classification method and device

Abstract

The text classification method and device disclosed in the present invention receive a plurality of training texts of known categories, preprocess the training texts, and construct the graph structure of the training texts using the co-occurrence relationship of words. According to the training The graph structure of the text, the parameters of the convolutional neural network are trained by the backpropagation algorithm to obtain the trained convolutional neural network, and then the input text to be classified is received, and the text to be classified is preprocessed Finally, the graph structure of the text to be classified is constructed using the co-occurrence relationship of words, and then according to the graph structure of the text to be classified, the category technical scheme of the text to be classified is predicted by the convolutional neural network after training, Apply convolutional neural network to the problem of text classification to improve the accuracy and credibility of text classification.

Description

technical field [0001] The present invention relates to the field of machine learning, in particular to a text classification method and device. Background technique [0002] Convolutional neural network is an artificial neural network with deep learning ability designed according to the principle of primate visual neural mechanism. Hubel and Wiesel proposed a visual structure model based on cat visual cortex in 1962, and proposed the concept of receptive field for the first time. However, with the emergence of simpler and more efficient linear classifiers such as support vector machines, and due to the local minimum limitations commonly found in non-convex objective cost functions of deep structures, the research on neural networks has fallen into a low ebb for nearly two decades. It is known that Hinton et al. proposed an unsupervised bead layer training method based on the deep belief network (DBN), which solved the optimization problem related to the deep structure. ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the embodiment of the present invention is to provide a text classification method and device, which can effectively solve the problem of lack of application of convolutional neural network for text classification in the prior art, and improve the accuracy and credibility of text classification

Images