TWDS Group20 Medical Project

Note!! This repository is currently deprecated and I have moved it to an organization to continue the development, please refer to our new repository

Pre-requisite

In this project, we use mlflow to track the records of each training experiments. Before you step into this project, please make sure your are familar with building a mlflow localhost server.

Introduction

Nowadays, many people feel unhappy in their daily life. Moreover, someone have suffered the depression for a while. We collect the Quetionnaires from the CDC of United States to explore what kind of people may be in the risk of depression.

Dataset

We selected the Questionnaires released annually from The CDC in United States to study our topic, due to its completeness. Each questionnaires contain 300 questions to answer, the questions contain either catebgorical or numeric information,

source of the dataset: https://www.cdc.gov/brfss/questionnaires/index.htm

Model Training

We provided two kinds of method to do model seletion:

1. Hold-Out Training
2. Cross-Validation

The corresponding command lines are as follows:

Train a model with hold-out validation

$ python train_model.py

Train a model with cross-validation

$ python train_model_cv.py

In both train_model.py, and train_model_cv.py, please setup the parameters to the experiement along with model parameters and our dedicated Training Framework will train the model and report the performance.