Skip to main content

Table 8 Percentage improvement of proposed models over the benchmark for Clinic 1 across all years

From: Forecasting patient flows with pandemic induced concept drift using explainable machine learning

 

2017

2018

2019

2020

2021

2022

Overall

Prophet

26.8

29.8

25.5

27.7

26.1

3.3

25.0

Random Forest

17.1

26.2

27.8

38.5

29.7

22.2

27.3

CatBoost

13.6

27.6

26.5

43.5

30.5

25.3

28.1

Gradient Boosting

18.4

26.9

28.5

43.1

31.3

24.6

29.2

Stacking

25.7

30.6

30.1

40.5

31.0

19.1

30.5

Averaging

24.6

31.3

32.1

42.0

34.5

25.4

32.2

Voting

25.5

31.0

32.3

41.2

36.1

27.2

32.7