jundelikwo / hospital-mortality-prediction

In Hospital Mortality Prediction

The predictors of in-hospital mortality for intensive care units (ICU)-admitted HF patients remain poorly characterized. We aimed to develop and validate a prediction model for all-cause in-hospital mortality among ICU-admitted HF patients.

This project attempts to create a classifier that can identify patients that will die or remain alive based on certain parameters.

Source

This project is an implementation of the kaggle project found at https://www.kaggle.com/saurabhshahane/in-hospital-mortality-prediction

Models

There are two models in this project

1. xgboostmodel.py

This uses the XGBClassifier. A summary of the performance of this model is given below

	precision	recall	f1-score	support
0 (Alive)	0.95	0.96	0.95	204
1 (Dead)	0.96	0.95	0.95	204
accuracy			0.95	408
macro avg	0.95	0.95	0.95	408
weighted avg	0.95	0.95	0.95	408

2. ann.py

This uses an artifical neural network. A summary of the performance of this model is given below

	precision	recall	f1-score	support
0 (Alive)	0.98	0.89	0.94	204
1 (Dead)	0.90	0.99	0.94	204
accuracy			0.94	408
macro avg	0.94	0.94	0.94	408
weighted avg	0.94	0.94	0.94	408