chenxichen95 / GP-epilepsy_qa_system

Epilepsy QA System

Data source

We collected epilepsy data by use web crawler from some website, the form of raw data as follow:

{
    "_id" : ObjectId("5b768fb956339877d3d947a3"),
    "illnessType" : "癫痫",
    "Q" : "老年人群体为什么外伤容易导致癫痫",
    "Q_detailed" : "外公患上癫痫之后，相知道老年人群体为什么外伤容易导致癫痫",
    "A1" : "任何年龄的脑外伤后都可能发生癫痫病，一般脑外伤后要服二年的抗癫痫病药物预防。\n      指导意见：\n      癫痫病一旦确诊，就需要严格按医生分咐，按时按量，长期服用抗癫痫病药物治疗，不能随便减量，更不能随便停药。"
}

Model

Our Q&A model mainly refer to DMN and DMN+.
Model frame as the following picture:

QA System Web Server

We use Flask to build a background server, and write a front-end page. Our model deployed on the background server. User can input their question, our system will return a answer.
The front-end page as the following picture:

Environment

• pymongo 3.7.1
• jieba 0.39
• numpy 1.14.5
• pandas 0.23.3
• gensim 3.5.0
• tqdm 4.25.0
• tensorflow 1.9.0
• Flask 1.0.2

How to run

if you want load and preprocess data, build and train word2vec (CBOW), build, train and test model, you can run the code as follow:

  sh run.sh

if you want start QA System Web Server, you can run the code as follow:

  modelPath='../model_build_train_dev_test/logs/2019-04-26@12_34_08-batchSize20-epoch10-embeddingSize150-dropout0.0-hidden_units80-hops3-lr0.001-randSeed1000/model/model-5'
  sh run_qa_system_web_server.sh -m modelPath