honeyandme/RAGQnASystem

• hightopics#1

  Add relevant topics to the repository

  Why:

  COPY-PASTE FIX

  Copy

  medical-qa, knowledge-graph, rag, llm, neo4j, bert, nlp, healthcare, artificial-intelligence, python, streamlit

  OpenDoneDismiss
• highreadme#2

  Refine README's opening to emphasize 'complete system' and 'medical' niche

  Why:

  CURRENT

  本项目构建了一个**基于知识图谱 RAG（Retrieval-Augmented Generation）+ 大语言模型**的医疗问答系统：
  - 以 **Neo4j 医疗知识图谱**（4.4 万实体、31 万关系）为外部知识源
  - 结合 **BERT + RNN** 做命名实体识别（NER），F1 达 97.40%
  - 利用 **ollama 本地 LLM** 做意图识别（16 类）与流式答案生成
  - 通过 **Streamlit** 提供完整交互界面
  
  区别于传统向量数据库 RAG，本项目使用**结构化知识图谱**做精确检索，为大模型提供可靠的医疗领域外部知识，有效缓解大模型在医疗场景的幻觉问题。

  COPY-PASTE FIX

  Copy

  This project implements a **complete medical intelligent Q&A system** leveraging **Knowledge Graph RAG (Retrieval-Augmented Generation) and Large Language Models**. Unlike traditional vector database RAG, it uses a **structured Neo4j medical knowledge graph** (44K entities, 310K relations) for precise retrieval, significantly mitigating LLM hallucinations in healthcare scenarios. It integrates **BERT + RNN** for Named Entity Recognition (F1 97.40%), **Ollama local LLMs** for intent recognition and streaming answer generation, and provides a full interactive UI via **Streamlit**.

  OpenDoneDismiss
• mediumlicense#3

  Add a LICENSE file to the repository

  Why:

  COPY-PASTE FIX

  Copy

  Create a `LICENSE` file in the repository root. Choose an appropriate open-source license (e.g., MIT, Apache-2.0, GPL-3.0) and add its text to the file.

  OpenDoneDismiss

Category GEO backends resolved for this scan: google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• CATEGORY QUERY

  How to build a reliable medical Q&A system preventing large language model hallucinations?

  you: not recommended
  AI recommended (in order):

  1. GCP Healthcare API
  2. Vertex AI
  3. Gemini
  4. Azure Health Data Services
  5. Azure OpenAI Service
  6. GPT-4
  7. Azure AI Search
  8. AWS HealthLake
  9. Amazon Bedrock
  10. Titan models
  11. AI21 Labs
  12. Anthropic
  13. Cohere
  14. Meta
  15. Stability AI
  16. Knowledge Bases for Amazon Bedrock
  17. Hugging Face Transformers
  18. Llama 2
  19. Mistral
  20. BioGPT
  21. LangChain
  22. LlamaIndex
  23. Pinecone
  24. Weaviate
  25. ChromaDB
  26. Med-PaLM 2
  27. OpenAI API
  28. GPT-3.5 Turbo
  29. Azure AI Content Safety
  30. Google Cloud's safety filters

  AI recommended 30 alternatives but never named honeyandme/RAGQnASystem. This is the gap to close.

  Show full AI answer
• CATEGORY QUERY

  Seeking a framework for accurate domain-specific question answering using knowledge graphs and LLMs.

  you: not recommended
  AI recommended (in order):

  1. LangChain
  2. LlamaIndex
  3. Haystack
  4. GraphRAG
  5. Neo4j
  6. Amazon Kendra

  AI recommended 6 alternatives but never named honeyandme/RAGQnASystem. This is the gap to close.

  Show full AI answer

• Metadata completeness

  fail

  Suggestion:

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of honeyandme/RAGQnASystem?

  pass
  ✓AI named honeyandme/RAGQnASystem explicitly

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?

• If a team adopts honeyandme/RAGQnASystem in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI named honeyandme/RAGQnASystem explicitly

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?

• In one sentence, what problem does the repo honeyandme/RAGQnASystem solve, and who is the primary audience?

  pass
  ⚠AI did not name honeyandme/RAGQnASystem — likely talking about a different project

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?