bitan1998/ann-training-using-diferential-evolution

Multiple randomized ANN are being generated that is being taken from user input(total number of ANN) then we have approached one of the nature-inspired-algorithms such as DIFFERENTIAL-EVOLUTION(DE) on a soil-content-dataset to prove that it has better prediction and optimising values other than some well defined algorithms such as SUPPORT-VECTOR-REGRESSION(SVR) and MLP-REGRESSOR.

FINAL-YEAR-PROJECT

ann-training-using-diferential-evolution's Issues

error

Got error

ValueError Traceback (most recent call last)
in
----> 1 f,b = de(bounds=[(-5,5)] * (3 * nodes_in_hidden_layer + nodes_in_hidden_layer) , popsize=number_of_ann, hidden_nodes=nodes_in_hidden_layer)
2 print(b)
3 plt.xlabel('ITERATION')
4 plt.ylabel('FITNESS')
5 plt.title('Evolution of fitness on 45235 iterations')

in de(bounds, popsize, hidden_nodes, mut, crossp, its)
13 for j in range(popsize):
14 idxs = [idx for idx in range(popsize) if idx != j]
---> 15 a, b, c = pop[np.random.choice(idxs, 3, replace = False)]
16 mutant = np.clip(a + mut * (b - c), 0, 1)
17 cross_points = np.random.rand(dimensions) < crossp

mtrand.pyx in mtrand.RandomState.choice()

ValueError: Cannot take a larger sample than population when 'replace=False'