to be or not to be ?
Please refer to the โCodaLab Challangeโ for more details.
To obtain the dataset:
Join the Competition & then:
wget https://competitions.codalab.org/my/datasets/download/9d49b18f-3861-44c1-8412-67fbb60dc0a3
The training dataset contains information about 80,000 patients, represented by categorical, binary and numerical features including age, gender, ethnicity, marital status, as well as medical data such as blood pressure rate or glucose rate. The following models were employed to predict if the patient died or not.
- LightGBM
- XGBClassifier
- CatBoost Cross Validation was performed using Optuna. This landed me on the 12th spot on the final leaderboard.