Table 8 Prediction results for \(^{13}C\) shifts using increasing numbers of spectra. n/a indicates that not enough data were available, numbers in brackets indicate number of compounds used, deviating from the top header

Chloroform-D1 (CDCl3)

2019 model

MAE (ppm)

58.62

50.91

24.06

RMSE (ppm)

74.63

66.18

34.38

MASE

1.21

10.4

0.48

\(\sigma\) (ppm)

52.01

48.61

29.47

2023 model

MAE (ppm)

5.23

4.32

4.12

RMSE (ppm)

12.42

10.62

10.53

MASE

0.10

0.08

0.08

\(\sigma\) (ppm)

12.17

10.47

10.48

HOSE code

MAE (ppm)

4.32

4.03

3.19

RMSE (ppm)

8.03

7.76

6.99

MASE

0.08

0.08

0.06

\(\sigma\) (ppm)

8.02

7.76

6.99

Missing predictions

17.30

19.42

30.25

Dimethylsulphoxide-D6 (DMSO-D6, C2D6SO)

2019 model

MAE (ppm)

85.84 (456)

n/a

n/a

RMSE (ppm)

93.76 (456)

n/a

n/a

MASE

1.92 (456)

n/a

n/a

\(\sigma\) (ppm)

45.01 (456)

n/a

n/a

2023 model

MAE (ppm)

5.75(456)

n/a

n/a

RMSE (ppm)

8.85 (456)

n/a

n/a

MASE

0.12 (456)

n/a

n/a

\(\sigma\) (ppm)

7.61 (456)

n/a

n/a

HOSE

code

MAE (ppm)

3.65 (456)

n/a

n/a

RMSE (ppm)

6.21 (456)

n/a

n/a

\(\sigma\) (ppm)

6.21 (456)

n/a

n/a

MASE

0.08 (456)

n/a

n/a

Missing predictions

10.0 (456)

n/a

n/a

Methanol-D4 (CD3OD)

2019 model

MAE (ppm)

69.41 (351)

n/a

n/a

RMSE (ppm)

82.03 (351)

n/a

n/a

MASE

1.60 (351)

n/a

n/a

\(\sigma\) (ppm)

49.75 (351)

n/a

n/a

2023 model

MAE (ppm)

5.53 (351)

n/a

n/a

RMSE (ppm)

9.94 (351)

n/a

n/a

MASE

0.12 (351)

n/a

n/a

\(\sigma\) (ppm)

9.11 (351)

n/a

n/a

HOSE code

MAE (ppm)

3.64 (351)

n/a

n/a

RMSE (ppm)

7.37 (351)

n/a

n/a

\(\sigma\) (ppm)

7.37 (351)

n/a

n/a

MASE

0.08 (351)

n/a

n/a

Missing predictions

12.5 (351)

n/a

n/a