onblog/Doctor

基于知识图谱的医学诊断系统。Medical Diagnosis System Based on Knowledge Map.

What it solves

This project addresses the difficulty of navigating vast amounts of fragmented medical data to provide accurate health assessments. It aims to help patients perform self-evaluations of their conditions and assist doctors in finding optimal treatment plans, thereby improving diagnostic efficiency and quality.

How it works

The system utilizes a medical knowledge graph to connect diseases and symptoms. It operates through the following process:

  1. Data Acquisition: Medical data is crawled from authoritative sources and stored in a structured format using MySQL, Elastic Search (for fast retrieval), and Neo4j (for graph-based analysis).
  2. Knowledge Graph Construction: The system creates nodes for symptoms and diseases. By segmenting the symptom attributes of diseases, it establishes direct links between them.
  3. Diagnostic Algorithm: When a user inputs symptoms, age, gender, and occupation, the system retrieves the relevant sub-graph from Neo4j. It then filters out impossible diseases based on demographic data (e.g., excluding male-specific diseases for female users) and calculates a "relevance score" based on the number of matching symptoms to rank potential diagnoses.
  4. Architecture: The system is built using a Java-based microservices architecture (Spring Cloud) to ensure high availability and scalability.

Who it’s for

  • Individuals: For those seeking quick self-diagnosis and a better understanding of their health conditions.
  • Medical Professionals: For researchers and doctors who need a tool to discover associations between symptoms and diseases to support their diagnostic process.

Highlights

  • Multi-Database Strategy: Combines MySQL, Elastic Search, and Neo4j to balance structured storage, fast search, and complex relationship analysis.
  • Demographic Filtering: Increases accuracy by using a specialized population distribution table to exclude irrelevant diseases based on user profile.
  • Relevance Scoring: Employs a two-step sorting process to rank diseases based on symptom match density.
  • Microservices Design: Uses a distributed architecture to handle high concurrency and maintainability.

Related

  • Project
    mozilla/translations

    Training pipelines and an inference engine for machine translation models that power local, client-side web page translation in the Firefox browser.

  • Project
    jxx123/simglucose

    simglucose is a Python implementation of the FDA‑approved UVa/Padova Type‑1 Diabetes simulator, packaged as a Gym‑compatible environment for reinforcement‑learning research. It offers 30 virtual patients, customizable reward functions, parallel execution, scenario generators, and built‑in visualisation, letting users test any insulin‑delivery controller (PID, MPC, RL, etc.) in a safe, reproducible setting.

  • Project
    djcopley/ShellOracle

    A terminal utility that generates shell commands from natural language descriptions and inserts them directly into the prompt using various LLM providers.

  • Project
    mvilanova/intervals-mcp-server

    An MCP server that connects Claude and ChatGPT to the Intervals.icu API, enabling AI assistants to retrieve and manage athletic training, wellness, and event data.