A biomedical event jointed extraction method based on a replication mechanism

Abstract

The invention provides a biomedical event extraction method based on a replication mechanism, belonging to the technical field of natural language processing. The biomedical event extraction method based on the replication mechanism comprises the steps of constructing a model input vector; constructing an Encoder module using a bi-directional LSTM model; constructing a Decoder module based on an Attention mechanism and a replication mechanism to recognize triggers and elements at the same time. The method of the invention can effectively avoid the disadvantages of cascading error and independence between subtasks caused by the staged method, and error transmission caused by the simple use of shared parameters in the joint model, thereby improving the performance of biomedical event extraction.

Description

technical field [0001] The invention belongs to the technical field of natural language processing, and relates to an event joint extraction method based on biomedical text. Specifically, it refers to using the joint extraction method through the replication mechanism to simultaneously extract trigger words and elements in biological events to form biological event candidates. Then classify the biological event candidates through the event structure features learned by the support vector machine (SVM), and remove the invalid combination pairs, so as to obtain a complete biomedical event. Background technique [0002] At present, there are two main types of methods to achieve biological event extraction. One is a staged method (also known as a pipelined method), which divides the extraction of biological events into two main steps: trigger word recognition and element recognition, and then constitutes a complete biomedical event through post-event processing. The other is t...

AI Technical Summary

Problems solved by technology

[0010] The staged method makes the biological event extraction task more hierarchical, but there are also some problems: (1) error propagation

Since the feature recognition stage needs to use the trigger words predicted in the trigger word recognition stage, if the trigger word recognition effect is not good, the error in the trigger word stage will be propagated to the feature recognition stage, resulting in cascading errors

(2) Neglecting the relationship between two subtasks

But this method still has some disadvantages: (1) When using the joint extraction method for biological event extraction, the extracted features depend on NLP preprocessing tools, which may cause errors

This method can reduce the independence between the two subtasks, but there is still the problem of error transmission

Images