how can i enter more than one input in NARX model?

Hi!

I want to train a NARX model, which have 1 target time series data and 3 input time series data. I tried to use the nnstool, but there I could use only 1 input data. How could I train my network with more than 1 input?

thanks, Roley

Answer 1

[Fran Mat]

Hi Roland:

You have to arrange your 3 input series data into 1.

Please check the example "pH Neutralization Process" that comes with the system.

Run these commands and check afterwards the variable "inputSeries". This one contains two series as input.

load pH_dataset

inputSeries = phInputs;

targetSeries = phTargets;

More info here:

http://www.mathworks.ch/ch/help/nnet/gs/neural-network-time-series-prediction-and-modeling.html

Best regards. Fran.

[Matheus Torquato]

What if I have a vector for each item from my input?

I have 200 cycles of data from 100 engines (with 24 parameters each).

{Engine1_Time001}    {Engine2_Time001}            {Engine200_Time001}
{Engine1_Time002}    {Engine2_Time002}   .  .  .  {Engine200_Time002}
{Engine1_Time003}    {Engine2_Time003}            {Engine200_Time003}
 .                    .                  .  
 .                    .                     .
 .                    .                        .
{Engine1_Time200}    {Engine2_Time200}            {Engine200_Time100}

Where {EngineX_TimeY} is a 1x24 Cell containing 24 parameters of the Engine X at cycle Y.

The Neural Network Time Series App doesn't seem to be able to handle that structure since the options available are only Cell Column, Matrix Column and Matrix Row.

How can I train the NARX then?

Answer 2

[Roland]

Thanks for your help,I see now how in that example are the inputs organized. Unfortunatelly I'm not so familiar with the matlab. Could you give any advice, or a link to a description to see how can I my imported data with more variable(from excel) reform for an usable Matlab input? for example:

8,8758 1 --->[8,8758;1]

thx, Roland

[Fran Mat]

Hi Roland:

Actually this is what I am looking for as well. I have to put 5 variables into one. If I find something, will let you know.

Best regards.

[Roland]

thanks!