Skip to main content
. 2018 Nov 23;15(12):2628. doi: 10.3390/ijerph15122628

Figure 4.

Figure 4

The comparison of the two periods (1950–2003, left; 1950–2015, right) between the measured and estimated value when using the Bayesian regulation back propagation neural network (BRBPNN) model for training (a), testing (b), and entire database (c). NS: Nash-Sutcliffe model efficiency.