What is the difference between trainlm and trainrp?

Hi all does the training function affect on the run data? I compared the result between two different functions (trainrp, trainlm). the obtained results are different. the accuracy resulted by trainlm is much better than trainrp (i.e. r value) but the concept behind the studied role appeared by trainrp is used!

thank you for help in advance.

Answer 1

[Greg Heath]

Before asking questions about any functions it is good practice to first use the MATLAB info via, e.g.,

 help trainrp
 doc  trainrp
 type trainrp

You will find reference to the terms

 resilient backpropagation algorithm and RPROP

which can be used in GOOGLE and WIKIPEDIA

Hope this helps.

Thank you for formally accepting my answer

Greg

[Lilya]

Dr. Heath, why trainlm is much better in classification rather than trainrp? (i.e. in pattern recognition)

[Greg Heath]

Why aren't you using the classification/patternrecognition net with TRAINSCG default ?

 help patternnet
 doc patternnet

 net = patternnet    % NO SEMICOLON!!!

Greg

[Lilya]

Sorry Dr, I mean trainlm in prediction. I used it as you told me (NARXNET). therefore, the question becomes why trainlm is much better in prediction rather than trainrp?

[Greg Heath]

I don't know. In 36 years of designing NNs, I don't believe I ever used trainrp.

Greg

[Lilya]

Thank you Prof for your reply.