I'm a newbie..I don't understand why we use threshold and bias in MLP (Multi-Layer Perceptron
)??What is the role of threshold, bias..And I don't know the formula of output (the result after we use the activation function, such as sigmoid) follow bias and threshold..In the same document, I see:
output_value=activation_function(summing_function+threshold) (follow Jeff Heaton)
output_value=activation_function(summing_function-threshold)(follow my teacher)
output_value=activation_function(summing_function+bias) (no problem!)
Which is the correct??Please give me a response!
And, the bias and the threshold can exists same time in MLP??