Supervised Training with Delta Rule

Asked By Senanayaka 0 points N/A Posted on - 09/30/2011

In a Neural Network there are two methods that we can train the network. One of method is Supervised Training. To do that Training we can use the Delta Rule. This is one of Example that I had found.

Can anybody explain me how to find the Error of this question?

• Status: Open
• Question Views: 791
• Answer Count: 1
• Vote Up 0 Vote Down

• Answer Accepted: No
• Question Category: Networking

Answered By akiltec 0 points N/A #84371

Supervised Training with Delta Rule

This is how we explain the supervised training of the Neural Network

Supervised Training

Training Data Set and Test Data Set are selected from Past or Historical Data.

Targets are selected

Weights are initially set randomly.

Input is applied and the Network Output is determined

Error = Target – Network Output is calculated

Weights are updated until Error à 0

 

This is the theoretical explanation of the Delta rule

Supervised Training – Delta Rule

Dw = hdx

Dw = wnew – wold

h = Learning Rate Parameter

d = Error (= Target – Network Output)

x = Input Vector

wnew – wold= hdx

wnew = hdx + wold

 

Let’s see the example that has given,

Lets assume the Target Is=1

So Network output, Y

Y = 1 x 0.3 + 0.5 x 0.4 + 0.3 x 0.5 = 0.65

Then we can find the Error

Error d = Target – Network Output = 1– 0.65 = 0.35

About akiltec

Questions
2

Answers
5

Best Answers
0

• Vote Up 0 Vote Down

• Posted on - 10/07/2011
• Question Category: Networking