how can Levenberg Marquardt algorithm be implemented in script to train a neural network

3 views (last 30 days)
Preeti Prajapati
Preeti Prajapati on 10 Jul 2014
Answered: Greg Heath on 12 Jul 2014
I have constructed the network to fit a time series data in script. The network has two hidden layers of size 4 each and want to train it using Levenberg Marquardt algorithm taking mse as the performance. I am seeking help as to how should I implement this algorithm in script as it is not very much clear in MATLAB help. I will be very grateful if you can explain it with a simple example as i am unable to find one. Thank you.

Accepted Answer

Greg Heath
Greg Heath on 12 Jul 2014
help narxnet doc narxnet
One hidden layer should be sufficient. However you should use the target autocorrelation and target-input crosscorrelation functions to help determine which input and feedback delays to use.
greg narxnet nncorr
Hope this helps.
Thank you for formally accepting my answer
Greg

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!