Derivar cinco veces la función sigmoide y otra funcion
The sigmoid is one of the earliest and simplest activation functions used in neural networks to introduce nonlinearity within each layer.
It has a useful property: its derivative can be expressed directly in terms of the function itself:
σ′(x) = σ(x)(1 − σ(x))
This makes it particularly convenient in backpropagation. However, the function quickly saturates toward 0 or 1 and in those regions the derivative becomes very small (with a maximum of 0.25 at x = 0) which slows down learning in deep networks (vanishing gradient).