Table 2 Improved accuracy of four models on seven training datasets compared with the baselines.

MLENET³¹

7.00

0.95

7.00

9.13

3.74

2.31

0.46

Lightweight MSRDN¹⁹

6.02

5.64

7.40

8.11

4.60

3.94

1.11

SCAE³⁰

8.12

5.49

1.26

0.94

0.91

1.25

\(-0.02\)

Joint model¹⁹

2.04

6.76

5.74

7.02

3.19

3.22

0.82