Table 2 Overview of baseline statistical and ML model performance metrics on test set (2021)

From: Machine learning-based forecasting of daily acute ischemic stroke admissions using weather data

  

Weather features only

Weather and calendar features

Model type

Model name

MAE

(Ncount/day)

MAPE

(%)

RMSE

(Ncount/day)

MAE

(Ncount/day)

MAPE

(%)

RMSE

(Ncount/day)

Baseline statistical

Poisson

1.72

58

2.73

1.69

57

2.68

GAM

1.44

56

1.64

1.41

56

1.62

Machine learning (ML)

SVR

1.28

52

1.58

1.27

52

1.58

RF

1.26

49

1.54

1.25

49

1.52

XGB

1.25

48

1.52

1.21

47

1.49

  1. GAM generalized additive model, MAE mean absolute error (Ncount/day), MAPE mean absolute percentage error (%), RMSE root mean square error (Ncount/day), RF random forest, SVR support vector regression using linear kernel, XGB extreme gradient boosting.