python - Predict Number using Neural network

Question

import numpy as np

def sigmoid(x):
    return 1/(1+np.exp(-x))

def sigmoid_derivative(x):
    return x*(1-x)


training_inputs=np.array([[1,2,3],
                          [4,5,6],
                          [7,8,9],
                          [6,7,8]])



training_outputs=np.array([[4,7,1,9]]).T

bias = np.array([[-10,-10,-10,-10]]).T
np.random.seed(1)
synaptic_weights=np.random.random((3,1))


for iteration in range(20000):
    input_layer= training_inputs
    outputs = sigmoid(np.dot(training_inputs,synaptic_weights))
    error = (training_outputs/10) - outputs
    adjustments = error*sigmoid_derivative(outputs)
    synaptic_weights+=np.dot(input_layer.T,adjustments)

new_output=(sigmoid(np.dot(training_inputs,synaptic_weights)))*10
given_input = np.array([1,2,3])

expected_output= (sigmoid(np.dot(given_input,synaptic_weights)))*10
print(expected_output)

I need to predict the output when i give any 3 input numbers as i tried giving 1,2,3 in last. But it does not give expected output and always give values around 9 as outputs. But 1,2,3 input and corresponding output also given as the training set. Can you please suggest how this model can use for predict values between 1 to 9

Answer 1

等待大神答复