Entity relationship extraction method, terminal equipment and computer readable storage medium

Abstract

The invention discloses an entity relationship extraction method, terminal equipment and a computer readable storage medium. The method comprises the following steps: manually extracting a plurality of binary entity pairs according with a preset entity relationship from an electronic medical record text database as seed instances; for each seed instance, searching a sentence comprising the seed instance in an electronic medical record text database, and extracting a feature vector of the sentence; clustering the seed instances based on the feature vectors; generating an extraction template corresponding to the cluster according to seed instances and feature vectors with sentences corresponding to the seed instances; extracting candidate instances from an electronic medical record text database by using the extraction template; calculating the confidence coefficient of each candidate instance according to the entity relationship between the candidate instance and the extraction template, so as to confirm whether to take the candidate instance as a new seed instance to carry out next iteration according to the confidence coefficient. According to the method, the accuracy of electronic medical record entity relation extraction can be greatly improved.

Description

technical field [0001] The invention belongs to the technical field of natural language processing, and in particular relates to a method for extracting entity relations from medical electronic medical records based on semi-supervision, a terminal device and a computer-readable storage medium. Background technique [0002] In this increasingly informatized and intelligent era, medical and health services are also constantly developing towards informatization and intelligentization, and medical electronic medical records have gradually begun to play an increasingly important role in the field of medical and health. Medical records (Medical Records) are records of medical activities such as examination, diagnosis, and treatment by medical staff on the occurrence, development, and outcome of patients' diseases. It is also a patient's medical and health file that summarizes, organizes, and comprehensively analyzes the collected data, and writes it according to the prescribed for...

AI Technical Summary

Problems solved by technology

However, the current experimental research has not yet obtained ideal extraction results, and its accuracy and recall have not improved significantly

[0008] Semi-supervised relationship extraction can utilize a large amount of unlabeled data, and only needs to manually annotate a small number of entity relationships. The accuracy rate of the weak supervision method for the relationship extraction of medical electronic medical records is not high, and the quality requirements for the initial relationship seeds are relatively high. Noise may be introduced to cause semantic drift, and the recall rate is low.

Images