Small sample text classification method and system based on semi-supervised learning

Abstract

The invention provides a semi-supervised learning small sample text classification method and system, and relates to the field of semi-supervised text classification, and the method comprises the steps: S1, obtaining a to-be-classified text; S2, inputting the to-be-classified text into a pre-trained lookup table, and mapping the to-be-classified text into text representation by the lookup table; S3, inputting the text representation into a multi-layer perceptron to obtain a text label, wherein the text label is used as a text classification result. According to the method, the text representation of the to-be-classified text is obtained through the lookup table, and then the multi-layer perceptron obtains the text tag as the text classification result according to the text representation, so a better classification effect can also be obtained under the condition of performing text classification on the text data with small data volume and incomplete data tags; manual labeling of a large amount of text data is not needed, manual labeling cost is saved, and the method can be flexibly applied to different scenes of various data quantities and data label labeling conditions.

Description

technical field [0001] The present invention relates to the field of semi-supervised text classification, and more specifically, relates to a small-sample text classification method and system of semi-supervised learning. Background technique [0002] With the development of technology, the amount of data on the Internet has increased exponentially. In the face of massive text, intelligent processing technology can save computing resources and improve processing efficiency. Text classification is the basic technology of information retrieval and mining, which plays a vital role in managing text data. [0003] In recent years, text classification has gradually changed from a shallow learning model to a deep deep learning model. Compared with shallow learning-based methods, deep learning methods avoid artificially designing rules and features, and automatically provide semantically meaningful representations for text mining. Therefore, most text classification research work...

AI Technical Summary

Problems solved by technology

[0004] However, the existing text classification methods based on deep learning need to collect enough training data and manually mark labels, which is time-consuming and labor-intensive

Moreover, for text classification of text data in some fields, it is necessary to collect special data sets, and it is difficult to make all data with labels

[0005] In the training method and text classification method of the text classification model disclosed in the prior art, by inputting the text samples under each task in multiple tasks into its corresponding private feature extractor and public feature extractor, the text samples under multiple different tasks The private feature extractor and classifier are trained at the same time to obtain the trained text classification model; however, this method cannot be trained under the conditions of small amount of data and incomplete data labels, and it is necessary to collect a large amount of training data and manually label labels, and the training cost high

Images