Skip to main content

Table 7 Contrasting MAPE accuracies between models using all features versus those generated by proxy-only features for Clinic 1

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

 

2017

2018

2019

2020

2021

2022

Benchmark

12.6

12.9

11.7

29.9

17.5

14.8

Voting (proxy-features only)

11.0

9.7

9.7

12.4

10.0

11.3

Voting (without proxy features)

8.4

8.5

7.7

13.7

10.3

9.2

Voting (all features)

8.4

8.2

7.5

11.6

8.9

9.7