Term extraction method and device based on deep learning network and storage medium

Abstract

The invention provides a term extraction method and device based on a deep learning network and a storage medium. The method comprises the following steps of labeling a term on a target text; carryingout the word segmentation on the labeled target text to obtain a word segmentation text and extracting keywords; according to the keywords, training the pre-established RNN deep learning network to obtain the term prediction model, and obtaining the term prediction result outputted by the term prediction model; according to the term prediction results and term tagging corresponding to the targettext, training the pre-established CNN depth learning network to obtain the term extraction model, and obtaining the term extraction results outputted from the term extraction model. The method of theinvention integrates the RNN and CNN deep learning network to form a deeper deep learning network. According to the extracted keywords and the terminology labeling result of the target text, the method of the invention predicts and extracts the terminology of the target text, which can effectively improve the extraction speed of the terminology and realize the extraction of the Chinese terminology of the massive text.

Description

technical field [0001] The present invention relates to the technical field of term extraction, in particular to a term extraction method, device and storage medium based on a deep learning network. Background technique [0002] A term represents a profession or a research direction in a field. Term extraction has research significance in the field of natural language processing, especially in machine translation and cross-language information retrieval. [0003] There are three traditional term extraction methods: rule-based methods, statistics-based methods, and machine learning-based methods. The above methods are mostly manual or non-manual based on the information in the corpus, and the rate of term extraction is low. For In today's era of information explosion, it is difficult to extract Chinese terms from massive texts only by manual or semi-manual methods. Contents of the invention [0004] Based on this, the present invention proposes a term extraction method, de...

AI Technical Summary

Problems solved by technology

[0003] There are three traditional term extraction methods: rule-based methods, statistical methods, and machine learning-based methods. The above-mentioned methods are mostly manual or non-manual based on the information in the corpus, and the rate of term extraction is low. For In today's era of information explosion, it is difficult to extract Chinese terms from massive texts by manual or semi-manual methods alone.

Images