Preprocessing laboratory tests results data by COVID-19 patients from Hospital Sirio Libanes. This repository contains one Jupyter notebook and one .csv file.
This work used original data from COVID-19 Data Sharing/BR FAPESP, available at https://repositoriodatasharingfapesp.uspdigital.usp.br/, referent to Sirio Libanes hospital from Sao Paulo, Brazil.
File containing filters steps to preprocessing. Patients were labeled according to the origin of laboraty tests, as shown below:
-
GROUP_0 - patients with exams coming only from the emergency room (NO_SEVERE);
-
GROUP_1 - patients with exams from the emergency room and hospitalization (NO_SEVERE);
-
GROUP_2 - patients with exams from the emergency room and ICU (SEVERE);
-
GROUP_3 - patients with exams from the emergency room, hospitalization and ICU (SEVERE).
Contains the data resulted from preprocessing described in the notebook. Next are the main characteristics of the output dataset:
- File name: sirio_aprendizado_v3.csv
- Missing rate: 13.2%
- Instances: 4320
- Classes: 2 (SEVERE and NO_SEVERE)
- Features categorical: 2
- Features continous: 2
If this methodology was useful, please cite as below:
@inproceedings{LOPES_2022, series={IJCIEOM 2022}, title={Analysis of COVID-19 severity prognosis using hospital data}, ISSN={2317-8000}, url={http://dx.doi.org/10.14488/ijcieom2022_abst_0026_37576}, DOI={10.14488/ijcieom2022_abst_0026_37576}, booktitle={International Joint Conference on Industrial Engineering and Operations Management Proceedings}, publisher={International Joint Conference on Industrial Engineering and Operations Management}, author={LOPES, FILIPE LOYOLA and GASPAR, ADRIANO SIMÕES and LORENA, ANA CAROLINA}, year={2022}, month=oct, collection={IJCIEOM 2022} }