An enhanced method of assisted triage

CN115565655BActive Publication Date: 2026-01-30UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202211233086.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-10-10
Publication Date
2026-01-30
Estimated Expiration
2042-10-10

AI Technical Summary

Technical Problem

Existing assisted diagnostic systems rely on the integrity of the database, cannot process symptom information that has not appeared in the database, and there is a large difference between patients' colloquial descriptions and medical terminology, resulting in low diagnostic accuracy, long patient waiting times, and heavy workload for doctors.

Method used

A three-layer prediction module is adopted: a clustering model, a body node symptom graph model, and a disease prediction model. A doctor-patient dialogue database is constructed using historical doctor-patient dialogue data. Cosine similarity and TF-IDF feature matching are used in combination with DQN neural network for comprehensive diagnosis to improve accuracy.

Benefits of technology

It improves the accuracy and scientific rigor of assisted diagnosis, reduces patient waiting time, alleviates doctors' workload, and enhances the ability to match patients' verbal descriptions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115565655B_ABST
    Figure CN115565655B_ABST
Patent Text Reader

Abstract

This invention belongs to the field of intelligent triage technology, and specifically relates to an enhanced auxiliary consultation method. The technical solution proposed in this invention includes three prediction modules. These modules are combined: a clustering model utilizes historical doctor-patient dialogue data, using the doctor's diagnosis as the pre-diagnosis result to improve credibility and scientific rigor; a body node symptom graph better matches the patient's verbal descriptions, enabling more accurate symptom information acquisition; a neural network-based disease prediction model can accept a wider range of input descriptions, providing broader coverage; and the results of the three modules are integrated, and a voting process is used to generate the final pre-diagnosis result, improving the accuracy and scientific rigor of the diagnosis.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Method and device for traditional Chinese medicine pathogenesis syndrome differentiation based on human body model, and storage medium

    CN109830299A

  • Triage method and device, computer device and medium

    CN110675944A

  • Disease auxiliary diagnosis system and device based on clustering and storage medium

    CN112002419A