A machine learning project where, we predict if a patient requries ICU or not, based on the test reports using Python and Scikit-Learn.
So, the dataset for this project is taken from kaggle.
Link for the dataset : https://www.kaggle.com/S%C3%ADrio-Libanes/covid19
The notebook to refer is the final.ipynb
Our aim was to find a patient who is suffering from COVID should be admited to an ICU or not.
The data for each patient is taken for a window size of 2 hours from the time they are admited to the hospital to 12 hours.
There were many missing values needed to be handled and the details regarding the handling of missing values was explained in the notebook.
There insights we could find as was been visulised as well and they explained in the notebook.
Few factors were about :
- Age
- Gender
- Type of diseases they are suffering
- Etc..
After the getting the insights we built few machine learning models like,
So, when seen in the notebook the accuracy for all models are less than 90% which is pretty bad . The next step for this project is to increase the accuracy, for which we need to tune our models or add more factors in our models.
Will be updating the notebook and the readme file soon after we increase the accuracy of our models.