license: mit | |
Submission for Hacklytics 2024 Datascience Hackthon | |
# Training | |
Random forest model trained on TEDS(TREATMENT EPISODE DATA SET | |
ADMISSIONS) dataset. Random search was employed to hyperparamter fine tune the model. | |
# Results | |
An accuracy of 74% was achieved with lots of potential to increase accuracy by fruther cleaning the input data set and feature engineering. | |
### Most important features | |
![Most important features](features.png) | |