A method and system for automatic conversion of medical rules to executable rule graphs

CN122158185APending Publication Date: 2026-06-05BEIJING HUIMEI CLOUD TECHNOLOGY CO LTD +1
View PDF 5 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING HUIMEI CLOUD TECHNOLOGY CO LTD
Filing Date
2026-05-09
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

Existing technologies for medical rule construction and execution suffer from problems such as high communication costs, high programming requirements, complex rule modification, inaccurate handling of vague expressions, difficulty in guaranteeing code quality, inability to make incremental modifications, and separation of visualization and execution.

Method used

By introducing structured graph editing operations as an intermediate representation layer, a five-stage pipeline processing of parsing, verification, standardization, assembly, and validation is designed. Combined with an ambiguity grading clarification mechanism and session persistence, the automatic conversion of natural language medicine rules into executable directed acyclic graph rule graphs is achieved.

Benefits of technology

It enables medical professionals to create and maintain complex rules directly without programming, improving the efficiency and accessibility of rule building, ensuring the correctness and reliability of the output rule graph, supporting efficient incremental editing, and unifying visualization and executability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122158185A_ABST
    Figure CN122158185A_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of medical informatization, and particularly relates to a medical rule to executable rule graph automatic conversion method and system, wherein the method comprises the following steps: acquiring a user input natural language medical rule and a current workflow state, the current workflow state comprising a session identifier, a working mode, an existing rule graph and a dialogue history; based on the current workflow state, calling a large language model to parse the natural language medical rule into a structured graph editing operation sequence, wherein when a rule is detected to be seriously unclear, a clarification operation is output, otherwise a graph editing operation sequence is output; performing structure legality verification on each graph editing operation in the graph editing operation sequence to obtain a verified graph editing operation sequence; and performing standardization processing on the verified graph editing operation sequence to obtain a standardized graph editing operation sequence.
Need to check novelty before this filing date? Find Prior Art