A multi-round inquiry method based on knowledge utility calculation

By constructing a disease knowledge graph and using a greedy algorithm to select efficient inquiry content, the problem of lack of dynamic adjustment of inquiry content in multi-round consultations is solved, realizing an efficient and controllable disease diagnosis process and improving diagnostic efficiency and reliability.

CN122290951APending Publication Date: 2026-06-26QINGDAO HARBIN INSTITUTE OF TECHNOLOGY (WEIHAI) +2
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-18
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing multi-round consultation methods lack the ability to dynamically adjust the content of the questions, making it difficult to quantify the contribution of different questions to the diagnosis of the disease. This leads to redundant questions or omission of key symptoms, affecting the efficiency and reliability of diagnosis.

Method used

A disease knowledge graph is constructed. Through diagnostic knowledge graph construction and preprocessing, initialization of consultation, extraction of candidate diagnostic subgraphs, calculation of candidate diagnostic feature utility functions, and selection by greedy algorithm, the efficient query content is dynamically selected to gradually narrow down the scope of diseases.

Benefits of technology

It achieves a structured, interpretable, and controllable multi-round consultation process, improving diagnostic efficiency and reliability, avoiding repeated questioning, and ensuring a balance between the accuracy of diagnostic results and interaction costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122290951A_ABST
    Figure CN122290951A_ABST
Patent Text Reader

Abstract

This invention discloses a multi-round consultation method based on knowledge utility calculation. The method includes the following steps: Step S1, construction and preprocessing of a diagnostic knowledge graph; Step S2, consultation initialization and determination of diagnostic feature states; Step S3, extraction of candidate diagnostic subgraphs based on diagnostic feature states; Step S4, calculation of the utility function of candidate diagnostic features; Step S5, symptom query selection based on a greedy algorithm; Step S6, multi-round consultation state update; Step S7, disease diagnosis termination determination and result generation. This invention, by constructing a diagnostic knowledge graph and combining symptom utility evaluation with a greedy strategy-based active query mechanism, can prioritize the acquisition of symptom information with high diagnostic value within a limited number of consultation rounds, thereby effectively narrowing the range of candidate diseases and improving the accuracy of disease diagnosis, while reducing the number of invalid consultations and improving the overall efficiency of the intelligent consultation system.
Need to check novelty before this filing date? Find Prior Art