Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

Thread Subject:
Trend or Future value prediction on the basis of historical data

Subject: Trend or Future value prediction on the basis of historical data

From: praveen

Date: 7 Aug, 2012 10:46:12

Message: 1 of 3

Dear ,

I have to predict the cpu utilization on the basis of past (historical ) data.

My Excel sheet 1st column (date and time) and 2nd column cpu value. using Matlab code I have to read and predict the trend for future value

Please suggest following..

1) after selecting time series tool , getting following option NARX, NAR and Nonlinear Input-Ouput please suggest which one I have to chose for prediction.

2) how I will set Target time series, defining the desisired output y(t).

My target should be as like that it contains future value with first column date and time and other should have predicted cpu value

Please suggest.... Thanks & Regards Praveen Kumar

Subject: Trend or Future value prediction on the basis of historical data

From: Greg Heath

Date: 8 Aug, 2012 12:39:17

Message: 2 of 3

"praveen " <pra.mit2004@gmail.com> wrote in message <jvqrlk$ovd$1@newscl01ah.mathworks.com>...
> Dear ,
>
> I have to predict the cpu utilization on the basis of past (historical ) data.
>
> My Excel sheet 1st column (date and time) and 2nd column cpu value. using Matlab code I have to read and predict the trend for future value
>
> Please suggest following..
>
> 1) after selecting time series tool , getting following option NARX, NAR and Nonlinear Input-Ouput please suggest which one I have to chose for prediction.
>
> 2) how I will set Target time series, defining the desisired output y(t).
>
> My target should be as like that it contains future value with first column date and time and other should have predicted cpu value
>
> Please suggest.... Thanks & Regards Praveen Kumar

1.You haven't specified the sampling time or the amount of data
2. Exclude NARX because the external cause of the output is missing.
3. Plot the three variables on the same plot. Anything interesting?
4. Plot the output vs each input with the correlation coefficient in the legend
5. Plot the crosscorrelation between the output and each input and determine
   if there might be significant lags between input ad output.
6. If there are significant I/O lags, use NAR. Otherwise use fitnet or newff.

Hope this helps.

Greg

Subject: Trend or Future value prediction on the basis of historical data

From: Greg Heath

Date: 6 Oct, 2013 13:51:07

Message: 3 of 3

"Greg Heath" <heath@alumni.brown.edu> wrote in message <jvtmll$qnp$1@newscl01ah.mathworks.com>...
> "praveen " <pra.mit2004@gmail.com> wrote in message <jvqrlk$ovd$1@newscl01ah.mathworks.com>...
> > Dear ,
> >
> > I have to predict the cpu utilization on the basis of past (historical ) data.
> >
> > My Excel sheet 1st column (date and time) and 2nd column cpu value. using Matlab code I have to read and predict the trend for future value
> >
> > Please suggest following..
> >
> > 1) after selecting time series tool , getting following option NARX, NAR and Nonlinear Input-Ouput please suggest which one I have to chose for prediction.
> >
> > 2) how I will set Target time series, defining the desisired output y(t).
> >
> > My target should be as like that it contains future value with first column date and time and other should have predicted cpu value
> >
> > Please suggest.... Thanks & Regards Praveen Kumar

Revision:

 1.You haven't specified the sampling time or the amount of data
 2. Plot the three variables, Day of week, time of day, and cpuutilization
 on the same plot. Anything interesting?
 3. Plot the target vs each input with the correlation coefficient in the legend.
 4. Plot the crosscorrelation between the target and each input and determine
    if there might be significant lags between input and output. If so, consider
    timeddelaynet and narxnet.
 5. Plot the autocorrelation function of the target and determine if there are
     any significant feedback lags. If so, forget about timedelaynet and concentrate
     on narxnet.
 6. See the explanations and examples in
      help narxnet
      doc narxnet
  7. If you have trouble with your design, search narx and narxnet in the NEWSGROUP
      and ANSWERS.
 8. For further problems, post relevant code and error messages.

Hope this helps.

Greg

Tags for this Thread

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Contact us