Sign in to follow this  

Noob-a-fied

This topic is 3744 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hey Im trying to understand the basics of back propagation. I'm noobing it up a lot here, ala VBscript and nothing is optimized for jack or...you know what. Here is what is happening Im putting in 8 inputs between 0 and 1, with 2 outputs also between 0 and 1. Random weights have been generated for all layers The weights going from The input to the hidden layer have been summed and multiplied by their randomly assigned weights (first time around) The activation sigmoid function is then ran on the summed weights*input of each neuron in the hidden layer, the value of the activation functions are kept in a seperate array, same series for the output neurons. This is where i start losing it... Next the error of the entire network is calculated by adding together the difference of output neuron 1 (after activation function)and the answer desired for it+the output of output neuron 2 and the desired answer for it. the error for each output neuron is then calculated with respect to the entire network by ....outputneuronafteractivation*(1-outputneuronafteractivation)*(desiredvalue-outputneuronafteractivation) The delta for the weight change is the Learningrate(eta?)*errorforoutputneuron*neuronbefore activation.... this is where i start really dropping the ball...well i drop it here and the math isnt so confusing as what goes where, can someone so an interation math and all, in lay terms just using number and no notation, i can understand it but after a lot of staring and googling...please no c++ thank you

Share this post


Link to post
Share on other sites

This topic is 3744 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this