Table 5 Compare the error evaluation indexes of different base learners and Stacking integrated models on the COVID-19 dataset.

From: Analysis and prediction of infectious diseases based on spatial visualization and machine learning

Type

Algorithm

Error evaluation indicators

MAE

RMSE

MAPE

Base-learner

ARIMA

0.679

0.854

0.312

ELM

0.482

0.615

0.338

SVR

0.261

0.296

0.328

Wavelet

0.310

0.416

0.341

RNN

0.460

0.570

0.340

SGDM-LSTM

0.393

0.402

0.321

Adam-LSTM

0.400

0.441

0.318

RMSProp-LSTM

0.327

0.407

0.295

Stacking

Model I

0.185

0.221

0.269

Model II

0.181

0.219

0.253

Model III

0.135

0.162

0.204

Model IV

0.143

0.179

0.212

Model V

0.172

0.214

0.253

  1. Note: In this table, the base learner of Model I is composed of RMSProp-LSTM, SVR, ELM, RNN and Wavelet. The base learner of Model II is composed of RMSProp-LSTM, SVR, ELM and Wavelet. The base learner of Model III is composed of RMSProp-LSTM, SVR, ELM and RNN. The base learner of Model IV is composed of ELM, Wavelet and SVR. The base learner of Model V is composed of RMSProp-LSTM, ELM and SVR.
Back to article page