Skip to main content

Advances, Systems and Applications

Table 5 Comparison of model accuracy. Best values for each model and metric are marked in bold print

From: Model-based cloud service deployment optimisation method for minimisation of application service operational cost

Metric

Linear Regression

Lasso

LARS

MARS

ANN

response time, MRS

MAE

243.39

476.62

181.11

181.47

221.05

MAPE (%)

9.33

19.82

5.79

5.61

6.24

RMSE

368.56

653.35

263.47

270.47

363.62

Adj. R2

0.9942

0.9817

0.9968

0.9969

0.9953

cost, MRS

MAE

142.98

113.68

99.00

106.25

124.91

MAPE (%)

22.82

13.47

6.33

15.19

10.66

RMSE

270.99

257.48

212.86

200.94

202.71

Adj. R2

0.9963

0.9967

0.9977

0.9980

0.9979

response time, video streaming

MAE

266.80

287.26

168.65

104.87

221.05

MAPE (%)

10.54

20.50

11.98

5.09

6.24

RMSE

495.25

401.45

226.01

149.86

363.62

Adj. R2

0.9588

0.9729

0.9914

0.9962

0.9867

cost, video streaming

MAE

331.24

234.18

123.59

72.77

531.55

MAPE (%)

10.98

5.88

3.84

2.02

17.26

RMSE

463.54

333.45

209.41

186.87

719.94

Adj. R2

0.9982

0.9991

0.9996

0.9997

0.9750