文本处理的方法、装置、存储介质及电子设备

By training a pre-defined recognition model for each type of entity relationship, the entity connection type in the target text is identified, which solves the problem of error accumulation in existing medical text relationship extraction and achieves higher accuracy and recall.

CN116166818BActive Publication Date: 2026-07-17SHENYANG NEUSOFT INTELLIGENT MEDICAL TECH RES INST
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHENYANG NEUSOFT INTELLIGENT MEDICAL TECH RES INST
Filing Date
2023-01-10
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing deep learning-based methods for extracting medical text relationships suffer from error accumulation, resulting in low accuracy of extracted text relationships.

Method used

For each entity relationship to be extracted, a corresponding preset recognition model is trained. The preset recognition model is used to identify the target entity connection type of each character pair in the target text and extract the target entity pair whose entity relationship is the first entity relationship.

Benefits of technology

It improves the accuracy and recall of relation extraction, showing a significant improvement compared to traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116166818B_ABST
    Figure CN116166818B_ABST
Patent Text Reader

Abstract

本公开涉及一种文本处理的方法、装置、存储介质及电子设备，可以获取待处理的目标文本中的多个字符对；针对预先训练得到的至少一个预设识别模型中的每个预设识别模型，通过所述预设识别模型识别每个所述字符对分别对应的目标实体连接类型，所述目标实体连接类型表征所述字符对中的字符在目标实体中的字符位置，所述目标实体的头部字符和 / 或尾部字符位于所述字符对中，不同的预设识别模型对应不同的实体关系，实体关系为目标文本的至少一个实体对中两个实体之间的关联关系；根据预设识别模型识别的所述目标实体连接类型，从目标文本中提取实体关系为第一实体关系的目标实体对，第一实体关系为与所述预设识别模型对应的实体关系。
Need to check novelty before this filing date? Find Prior Art