Quality control method and system for medical text information generated by natural language processing model

CN122047229BActive Publication Date: 2026-06-23BEIJING YIYONG TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BEIJING YIYONG TECH CO LTD
Filing Date
2026-04-16
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

In existing technologies, medical text information generated by natural language processing models inevitably contains errors and uncertainties. Relying on manual quality control review is time-consuming and labor-intensive, and it is difficult to maintain consistent quality, which has become a bottleneck for large-scale clinical data research.

Method used

Multi-dimensional quality inspection is performed using a large language model, including uncertainty quality scoring, abnormal input detection, logical consistency and reasonableness scoring. Errors in the structured data output by NLP are automatically identified and verified, and a comprehensive quality control score is generated to assist manual review.

Benefits of technology

It significantly reduces the workload of manual review, improves the accuracy and consistency of structured data, reduces the probability of missing erroneous results, and realizes an efficient and reliable quality control method and system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122047229B_ABST
    Figure CN122047229B_ABST
Patent Text Reader

Abstract

The application provides a quality control method and system for medical text information generated by a natural language processing model. The method includes generating structured data of medical text information based on a natural language processing model; inputting the structured data and its corresponding original medical text into a large language model configured to determine multiple quality scores of the structured data, including: evaluating the degree of support of the original medical text for the structured data to determine an uncertainty quality score; determining an abnormal input detection quality score based on a complexity classification of the original medical text; evaluating the logical consistency of a second medical text associated with the original medical text with the structured data to determine a logical consistency quality score based on the logical consistency; and evaluating whether the structured data conforms to the rationality of medical logic based on a pre-constructed medical knowledge base to determine a rationality quality score; and generating a comprehensive quality control score for the medical text information according to the multiple quality scores of the structured data.
Need to check novelty before this filing date? Find Prior Art