I'm working on a neural network simulator, writing it in c++ using microsoft visual studio 6.
I've come to a road block.
i've defined a neuron class and a group class.
a neuron object holds all the information spesific to the individual neuron. like it's current output.
a group object holds a vector of neurons and information about the connections made by the neurons it contains and information about how they should process their output.
the connections between nodes are done directly with pointers.
in some situations i'm going to have to access information held by the group through a pointer to a neuron in the group.
how can I do that without giving each neuron a pointer to the group that it's contained in? i can't figre it out.
the best I can figure so far is to decompose the neuron class and have insted of a vector of neurons have a vector of neuron outputs, a vector of wieght vectors. but then the problem is that I the connectors need a pointer to a group and an index for the vector of outputs...
am I making sense?
any ideas?
thankyou
I've come to a road block.
i've defined a neuron class and a group class.
a neuron object holds all the information spesific to the individual neuron. like it's current output.
a group object holds a vector of neurons and information about the connections made by the neurons it contains and information about how they should process their output.
the connections between nodes are done directly with pointers.
in some situations i'm going to have to access information held by the group through a pointer to a neuron in the group.
how can I do that without giving each neuron a pointer to the group that it's contained in? i can't figre it out.
the best I can figure so far is to decompose the neuron class and have insted of a vector of neurons have a vector of neuron outputs, a vector of wieght vectors. but then the problem is that I the connectors need a pointer to a group and an index for the vector of outputs...
am I making sense?
any ideas?
thankyou