import numpy as np
X = np.array(([2,9], [1,5], [3,6]), dtype = float)
y = np.array(([.92], [.86], [.89]), dtype = float)
X = X/np.amax(X, axis = 0)
def sigmoid(x):
return 1/(1+np.exp(-x))
def der_sigmoid(x):
return x*(1-x)
epoch = 5000
lr = 0.01
neurons_i = 2
neurons_h = 3
neurons_o = 1
suni
weight_h = np.random.uniform(size=(neurons_i, neurons_h))
bias_h = np.random.uniform(size=(1, neurons_h))
weight_o = np.random.uniform(size=(neurons_h, neurons_o))
bias_o = np.random.uniform(size=(1, neurons_o))
for i in range(epoch):
inp_h = np.dot(X, weight_h) + bias_h
out_h = sigmoid(inp_h)
inp_o = np.dot(out_h, weight_o) + bias_o
out_o = sigmoid(inp_o)
err_o = y - out_o
grad_o = der_sigmoid(out_o)
delta_o = err_o * grad_o
err_h = delta_o.dot(weight_o.T)
grad_h = der_sigmoid(out_h)
delta_h = err_h * grad_h
weight_o += out_h.T.dot(delta_o) * lr
weight_h += X.T.dot(delta_h) * lr
print('Input:n', X)
print('Actual:n', y)
print('Predicted:n', out_o)
{"html5":"htmlmixed","css":"css","javascript":"javascript","php":"php","python":"python","ruby":"ruby","lua":"text\/x-lua","bash":"text\/x-sh","go":"go","c":"text\/x-csrc","cpp":"text\/x-c++src","diff":"diff","latex":"stex","sql":"sql","xml":"xml","apl":"apl","asterisk":"asterisk","c_loadrunner":"text\/x-csrc","c_mac":"text\/x-csrc","coffeescript":"text\/x-coffeescript","csharp":"text\/x-csharp","d":"d","ecmascript":"javascript","erlang":"erlang","groovy":"text\/x-groovy","haskell":"text\/x-haskell","haxe":"text\/x-haxe","html4strict":"htmlmixed","java":"text\/x-java","java5":"text\/x-java","jquery":"javascript","mirc":"mirc","mysql":"sql","ocaml":"text\/x-ocaml","pascal":"text\/x-pascal","perl":"perl","perl6":"perl","plsql":"sql","properties":"text\/x-properties","q":"text\/x-q","scala":"scala","scheme":"text\/x-scheme","tcl":"text\/x-tcl","vb":"text\/x-vb","verilog":"text\/x-verilog","yaml":"text\/x-yaml","z80":"text\/x-z80"}