Figure 18. How classification error changes with different no. of hidden nodes,  Elman recurrent, 2 segments, 30 inputs

Figure 19. How classification error changes with no. of input nodes,  one hidden layer