Text classification method and device, terminal and storage medium

Abstract

The invention discloses a text classification method and device based on a neural network model, a terminal and a storage medium, and the method comprises the steps: searching a learning rate according to the gradient of an optimizer based on a random gradient descent and momentum method in the deep learning training process of the text classification model, and then improving the attenuation ratein the optimizer according to the reduction ratio of the learning rate; Based on the improved attenuation rate, the searched learning rate and the gradient, optimizing the neural network model through an optimizer until the model completes optimization training; And finally, inputting the text word vector corresponding to the target text into the neural network model subjected to optimization training to obtain a text classification result. In the text classification model learning and training process, a neural network model is optimized by combining a momentum method, random gradient descent and dynamic adjustment of the learning rate and the attenuation rate in an optimizer, the convergence speed and the execution effect of the model are helped to be increased, and then the accuracy ofa classification result during text classification is improved.

Description

technical field [0001] The present invention relates to the field of computers, in particular to a neural network model-based text classification method, device, terminal and computer-readable storage medium. Background technique [0002] The current text classification model based on deep learning is to input the word vector into the trained neural network model to obtain the classification result of the word vector. In order to make the classification results more realistic, it is necessary to train the neural network model before text classification, and the most important link in the model training process is to optimize the neural network model through the optimizer. However, at this stage, the optimization method of deep learning models based on text classification mostly adopts the method of SGD (Stochastic Gradient Descent, stochastic gradient descent), which can achieve good accuracy, that is, low error rate, to a certain extent when the model convergence speed is i...

AI Technical Summary

Problems solved by technology

[0003] The main purpose of the present invention is to provide a text classification method, device, terminal and computer-readable storage medium based on a neural network model, aiming at solving the problem of inaccurate classification results when classifying text

Images