Mortality prediction in ICU Using a Stacked Ensemble Model

Algorithm 1

APACHE score-based model using stacked ensemble technique.

1: Input: Feature data

2: Number of k-fold cross validation K

3: Number of base learners M

4: New training data

5: Output:

6: Begin:

7: Layer 1: Construct base learners

8: Use features AHDP and AIDP to train the APACHE learner AL.

9: Divide the into training data and test data

10: for i in 1 to M do

11: for j in 1 to K do

12:

13: Generate the prediction based on the M base learners

14: end for

15: end for

16: Layer 2: Construct final learner

17: Generate the new training

18: Treat the results of AL as one of the

19: LR model is employed to learn the final model based on the new training set

20: Test on the original test set

21: Generate

22: return