Figure 10
The horizontal axis represents “The number of hidden layer neurons,” ranging from 4 to 16 in increments of 2, and the vertical axis represents “Mean Squared Error (M S E) (10 to the negative 3 power),” ranging from 2 to 9 in increments of 1. The line graph displays four curves. The S C G underscore Training curve (dashed line with square markers) starts at (4, 6.8), passes through (6, 5.0), (10, 3.7), and (12, 3.2), and ends at (16, 3.3). The S C G underscore Validation curve (dashed line with circle markers) starts at (4, 8.2), passes through (6, 4.2), (10, 4.0), and (12, 3.3), and ends at (16, 3.7). The L M underscore training curve (dashed line with triangle markers) starts at (4, 2.8), passes through (6, 2.7), (10, 2.75), and (12, 3.2), and ends at (16, 2.7). The L M underscore Validation curve (dashed line with inverted triangle markers) starts at (4, 4.9), passes through (6, 3.4), (10, 3.0), and (12, 3.3), and ends at (16, 3.2). Note: All numerical data values are approximated.Comparison on MSE of networks with LM and SCG algorithms under different hidden-layer size