Medical text relation extraction method based on pre-training model and fine tuning technology

Abstract

The invention relates to a medical text relation extraction method based on a pre-training model and a fine tuning technology. The method comprises the steps of preprocessing of medical relation extraction corpora, model pre-training and fine tuning. According to the method, the pre-training model is used as the input of the one-dimensional convolutional neural network model, but the word embedding is used as the input of the one-dimensional convolutional neural network model in the prior art, and the pre-training model is more favorable for improving the extraction performance of the medicaltext relationship compared with the word embedding; according to the method, the one-dimensional convolutional neural network model and the pre-training model are combined for use, and the one-dimensional convolutional neural network is used for finely adjusting the pre-training model, so that the performance of the model is improved; the training error of the one-dimensional convolutional neuralnetwork is propagated back to a pre-training model to realize a model fine tuning process which is a dynamic model training process; in a traditional method, word embedding is combined with input of different layers, a main task model is still trained from the beginning, pre-trained embedding is regarded as a fixed parameter, and the usability of the method is limited.

Description

technical field [0001] The present invention relates to the technical field of natural language data analysis and processing, in particular to a method for extracting text relationships in the medical field, and more specifically to a method for extracting medical text relationships based on pre-training models and fine-tuning techniques. Background technique [0002] There are two Chinese explanations about medical treatment: 1. Healing, 2. Treatment of diseases. The history of Chinese medicine has been around for thousands of years, but this word has only appeared in recent decades. It is a new word to be in line with international standards. Before that, treatment was mostly used, and medical care also included health care content. [0003] Medical data is of great value, especially in relation to medical texts. Medical data mining has risen to a national strategy, and it is also a research hotspot in the competition between academia and industry around the world. How ...

AI Technical Summary

Problems solved by technology

However, such methods generally require a large training corpus to achieve high performance.

Images