Cascade-forward neural network

I was trying to find purpose of cascade-forward neural network and didn't find anyresonable sources. Can anyone explain advantages and disadvantages of using this type of network instead of classic feedforward neural network?

Answer 1

[Greg Heath]

You are right.

Despite the name, these are not the same network.

Greg

[Ekta Prashnani]

So, what is exactly the purpose of the cascade network? Its not the same as cascade correlation network and it seems to be distinct from classic feedforward network as well. Can you please explain?

Answer 2

[Greg Heath]

If you had searched GOOGLE with

 CASCADE CORRELATION NEURAL NETWORK

YOU WOULD HAVE FOUND

 http://scholar.google.com/scholar?q=cascade+correlation+
neural+network&hl=en&as_sdt=0&as_vis=1&oi=scholart&sa=
X&sqi=2&ved=0ahUKEwjr2MSEpcrPAhWKNT4KHfAECx8QgQMIGjAA

Hope this helps.

THANK YOU FOR FORMALLY ACCEPTING MY ANSWER

Greg

[EdWood]

I had searched Google and found Cascade correlation network, but I wasn't sure if it's the same as Cascade-forward network in MATLAB. For Cascade correlation network I found this definition: We add hidden units to the network one by one. Each new unit therefore adds a new one-unit "layer" to the network, unless some of its incoming weights happen to be zero. This is confused me beacuse when I run Cascade network in MATLAB it was look like there is 10 neurons from beggining for example and I wasn't sure if it's the maximum of neurons only and it's really adding one by one during learning or there is 10 fixed neurons from beggining. If you say it's Cascade correlation network then it shoud be add one by one, right?

[Christoph Thale]

You are right. The Cascade Correlation is a method where we add perceptrons one by one. So we start very small and end up bigger. We stop there, where the net performs best.