SEMINAR ON BACKPROPOGATION : Backpropagation is a systematic method for training multilayer artificial neural networks. Has a mathematical basis that is stronger if it is not very practical. Despite its limitations, backpropagation has significantly broadened the range of problems that the RNA can be applied, and has generated many successful demonstrations of its power.

The neuron used as a fundamental element for backpropagation networks. A set of inpuits applies, either from abroad or from a previous layer. Each is multiplied by a weight, and products are summed. the sum of the products known as NET and should be calculated for each neuron in the network. NET is calculated after an activation function F is applied to the change, which produces the output signal.

Backpropagation network training:

1. Select the next pair of training throughout the training: using the input vector to the input of the network.

2. Calculate the output of the network.

3. Calculate the error between the output of the network and desired output (target vector pair formation).

4. Set the weight of the network in a way that minimizes the error.

5. Repeat steps 1 through 4 for each vector of the training, so that the error for the set is sufficiently low.

The neuron used as a fundamental element for backpropagation networks. A set of inpuits applies, either from abroad or from a previous layer. Each is multiplied by a weight, and products are summed. the sum of the products known as NET and should be calculated for each neuron in the network. NET is calculated after an activation function F is applied to the change, which produces the output signal.

Backpropagation network training:

1. Select the next pair of training throughout the training: using the input vector to the input of the network.

2. Calculate the output of the network.

3. Calculate the error between the output of the network and desired output (target vector pair formation).

4. Set the weight of the network in a way that minimizes the error.

5. Repeat steps 1 through 4 for each vector of the training, so that the error for the set is sufficiently low.

## No comments:

## Post a Comment