I've written a simple neural network learning with back propagation
algorithm.
My network has 6 inputs - these are large numbers (first 5 are between 0 and
540 and the last is between -180 and 180). It is a 3 layer network - 20
neurons in the input layer ,10 in hidden layer and 1 in output layer. I want
my network to return values from (0,1) than I wrote a function f: <0,1> ->
<-5,5> anf f^-1 because these are the values I need in the end.
The input layer has activation function like that: 1 / (1 + e^(-s/100)
hidden layer function is 1 / (1 + e^(-s)
output layer: 1 / (1 + e^(-s/10)
I'm training the network with a correct value which is always 0 (for test
purposes) but as the network back-propagates the output is going towards 1
(which is 5 in the end).
Can somebody figure out what may be the problem? If you need any more
information about the network just tell me.
Re: Need a script to understand names posts ... the physical layer of the ISO model, augmented by providing all of the logic ... a network model that does more than just detect the fact that packets ... and brings them back onto the network as if they had never been lost.... the packets previously lost will never ever be lost again, ... (microsoft.public.scripting.wsh)
Re: Machine learning ... They usually consist of neurons,...input layer, the intermediate layer and the output layer. ... into the network using a method called "back-propagation". ... Train the network until it recognises characters reliably..... (borland.public.delphi.thirdpartytools.general)
Re: Client Information ... simply collecting data of who is using it and when. ... I have created a WebService in VS 2005 using C# that returns the ... IP addresses belong to the Network... This means that the Network layer can do whatever it wants to with ... (microsoft.public.dotnet.framework.webservices)
Re: Second IP address ... > If I add multiple IP address on the same network card, ... > increase load over IP stack?... stack in the machine, the network services.... handles IP and that is on the 3rd layer.... (microsoft.public.win2000.networking)
Re: [kde] smb:// ... ODF or GIMP file on our network and double-click on it - GIMP and ... It will provide abstraction from network layer of the OS and support of custom ...Copying of a file into the temporary directory is sometimes problematical. ... (KDE)
