This code uses pretrained recurrent neural network models for normalizing free-text descriptions of causes of death into icd-10 codes. Code written in python 2.7

Installing required libraries:

pip install -r requirements.txt

After installing libraries nltk punkt model must be downloaded in python interpreter: import nltk nltk.download('punkt')

Unpack dictionary_vectors.zip

Usage:

python predict.py -c file_path
file_path is a path to the file with death certificates in CLEF eHealth Task 1 format

Note: For additional scripts used for training RNNs please contact [email protected]

Citing:

KFU at CLEF eHealth 2017 Task 1: ICD-10 Coding of English Death Certificates with Recurrent Neural Networks Z Miftakhutdinov, E Tutubalina - 2017

http://ceur-ws.org/Vol-1866/paper_64.pdf

BibTex:

@inproceedings{
    miftakhutdinov2017kfu,
    title={KFU at CLEF eHealth 2017 Task 1: ICD-10 Coding of English Death Certificates with Recurrent Neural Networks},
    author={Miftakhutdinov, Z and Tutubalina, Elena},
    year={2017},
    organization={CLEF}
}