NOR-cleaner

AI model for resolving duplicities in Czech National Oncology Register. Bachelor's thesis about this project is available here.

Table of contents

• Requirements
• Installation
  • Windows
  • Linux
• How to use
  • Edit paths
  • Run the scripts
    • Order of execution

Requirements

• Python $\ge 3.10$

Installation

Install the package either by:

• Downloading the repository as a ZIP file by clicking the on the badge at the beginning of this README file.
• Cloning the repository

Windows

Run the file install.ps1 either by right-clicking and selecting Run with PowerShell or by running the command in PowerShell:

.\install.ps1

Linux

Create a virtual environment and activate it (optional but recommended):

python -m venv venv

Activate the virtual environment:

source venv/bin/activate

Install the requirements:

pip install .

How to use

Edit paths

Edit the paths in the scripts/constants.py file, e.g. paths to the data or model.

Run the scripts

In terminal, run the command:

nor-cleaner [-h] {prepare,train,predict,evaluate} ...

Order of execution

1. Prepare the data for training.

nor-cleaner prepare

2. Train the model.

nor-cleaner train

3. Predict whether to preserve or drop a record.

nor-cleaner predict

4. (Optional) Evaluate the model using cross-validation on the training data.

nor-cleaner evaluate

NOTE: use the constants file in scripts/constants.py to set paths.