I have some back propagation neural network questions.?

Question

first: How does summing weights*values for each neuron give you data?

second: I have a hard time understandng what runnign the output through the sigmoid does as all my values coming from the input layer to the hidden layer are going to be between 0 and 15 (amount of neurons)...

I hope you can understand the question

Answer 1

The output of a node is typically output = f(sum over weighted input values), where f is typically the sigmoid function. The input values are between 0 and 1. The sigmoid is used to force the output also in between those values so that it can serve as input in the next layer.

Any function with a range between 0 and 1 would have done, but the sigmoid is used, because it's derivative f' has the property that  f'=f*(1-f) which facilitates the backward propagation calculations.