Skip to Main Content
Table 3

Accuracy of BatNN trained with various fmax, based on HadCM3, ECHAM5 and HadGEM3-RA

fmaxHadCM3
ECHAM5
HadGEM3-RA
rRMSEMAEErRMSEMAEErRMSEMAEE
0.91 2.05 1.72 0.79 0.93 1.53 1.65 0.83 0.94 2.10 1.66 0.76 
0.93 1.99 1.65 0.81 0.95 1.47 1.30 0.85 0.94 2.06 1.59 0.80 
0.96 1.74 1.50 0.84 0.97 1.38 1.19 0.88 0.96 1.95 1.51 0.80 
0.96 1.74 1.57 0.84 0.97 1.43 1.25 0.86 0.95 1.98 1.61 0.78 
0.95 1.97 1.66 0.81 0.94 1.51 1.59 0.82 0.93 2.20 1.88 0.74 
10 0.92 2.19 1.79 0.79 0.94 1.52 1.60 0.82 0.93 2.21 1.90 0.74 
fmaxHadCM3
ECHAM5
HadGEM3-RA
rRMSEMAEErRMSEMAEErRMSEMAEE
0.91 2.05 1.72 0.79 0.93 1.53 1.65 0.83 0.94 2.10 1.66 0.76 
0.93 1.99 1.65 0.81 0.95 1.47 1.30 0.85 0.94 2.06 1.59 0.80 
0.96 1.74 1.50 0.84 0.97 1.38 1.19 0.88 0.96 1.95 1.51 0.80 
0.96 1.74 1.57 0.84 0.97 1.43 1.25 0.86 0.95 1.98 1.61 0.78 
0.95 1.97 1.66 0.81 0.94 1.51 1.59 0.82 0.93 2.20 1.88 0.74 
10 0.92 2.19 1.79 0.79 0.94 1.52 1.60 0.82 0.93 2.21 1.90 0.74 

Note: Bold numbers are optimal fmax values obtained.

Close Modal

or Create an Account

Close Modal
Close Modal