AForge.NET

  :: AForge.NET Framework :: Articles :: Forums ::

Resul Expression of the Time Series Prediction with GP

The forum is to discuss topics from different artificial intelligence areas, like neural networks, genetic algorithms, machine learning, etc.

Resul Expression of the Time Series Prediction with GP

Postby vincenzodentamaro » Mon Feb 27, 2012 4:34 pm

Hello everyone and thank you for this excellent framework.
I am a beginner of Genetic Programming and I saw the example called Time Series Prediction with GEP/GP.
Now I see that there is a solution that comes out after the training fase.
What kind of format is that solution? can I simply have a function of the kinf: y = f(x); with x from 0.....n and maybe create a curve to draw.
Thank you in advance.

Vincenzo
vincenzodentamaro
 
Posts: 9
Joined: Mon Feb 27, 2012 4:28 pm

Re: Resul Expression of the Time Series Prediction with GP

Postby andrew.kirillov » Mon Feb 27, 2012 4:47 pm

Hello,

This may give some info.
With best regards,
Andrew


Interested in supporting AForge.NET Framework?
User avatar
andrew.kirillov
Site Admin, AForge.NET Developer
 
Posts: 3446
Joined: Fri Jan 23, 2009 9:12 am
Location: UK

Re: Resul Expression of the Time Series Prediction with GP

Postby vincenzodentamaro » Tue Feb 28, 2012 3:38 pm

thank you for your quick reply!
I am investigating to use your framework in the company where I work.

Thank you again!
vincenzodentamaro
 
Posts: 9
Joined: Mon Feb 27, 2012 4:28 pm

Re: Resul Expression of the Time Series Prediction with GP

Postby vincenzodentamaro » Mon Mar 05, 2012 2:59 pm

Sorry but $0 refers to what?
Thank you in advance
vincenzodentamaro
 
Posts: 9
Joined: Mon Feb 27, 2012 4:28 pm

Re: Resul Expression of the Time Series Prediction with GP

Postby andrew.kirillov » Mon Mar 05, 2012 3:25 pm

Maybe x[t-1]? Not sure what you really do to say more. What about reading documentation?
With best regards,
Andrew


Interested in supporting AForge.NET Framework?
User avatar
andrew.kirillov
Site Admin, AForge.NET Developer
 
Posts: 3446
Joined: Fri Jan 23, 2009 9:12 am
Location: UK

Re: Resul Expression of the Time Series Prediction with GP

Postby dobdo99 » Mon Mar 05, 2012 8:11 pm

Hi all, First post hope its not too negative for a newbie:

I downloaded and got the Aforge lib working, really good start for a newbie in C#, and its promises much!!!!

so a time series prediction: What I expect: given a set of data, it trains for best solution, and returns it!

Given a set of data values in data[] array, ok, mine is really simple because there is much complexity here and terms that are not explained:

double[] data = new double[10] { 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32};

code: returns solution of: $6 $0 +
which means ( If I have understood rpn, newbie on this too )
$6 + $0, advanced for a newbie :-)

Now lets get really clear on this:
where does it say what $6 and $0 relate too, constants, data, args of the function?

Andrew, please, just post a link to the actual page that EXPLAINS WHAT THESE ARE? the page you posted to the guy who reply here was not an answer, it confused him and me.

Given this Solution how do we apply it to get the next in the sequence, which I think must be 34, right?

Real simple, please cos i'm real stupid :-)

Kind Regards, Carl
dobdo99
 
Posts: 4
Joined: Mon Mar 05, 2012 7:51 pm

Re: Resul Expression of the Time Series Prediction with GP

Postby andrew.kirillov » Mon Mar 05, 2012 8:29 pm

dobdo99 wrote:code: returns solution of: $6 $0 +

The problem is that you don't try reading documentation and sample code. The documentation clearly says that $0 is x[t-1]. Now, if you have window size 5, then starting from $5 you have constants, which means that $6 is a second constant. Taking a very quick look at provided sample application, you can see that second constant is 2. So your expression is "x[t-1] + 2". This means that every new value of time series equals to previous one increased by 2. I think this statement is clear from the data sample you gave.

dobdo99 wrote:cos i'm real stupid :-)

If it is so, then start learning and use your brain.
With best regards,
Andrew


Interested in supporting AForge.NET Framework?
User avatar
andrew.kirillov
Site Admin, AForge.NET Developer
 
Posts: 3446
Joined: Fri Jan 23, 2009 9:12 am
Location: UK

Re: Resul Expression of the Time Series Prediction with GP

Postby dobdo99 » Mon Mar 05, 2012 9:34 pm

Cheers Andrew

Now if you could please just post the link to the exact page of documentation that makes this clear?

Because I asked a question and you seem to have failed to be able to read and or understand it and have failed to give me an answer?

Unless of course it does not exist? and you tell people to read the documentation as if it explains everything?

Its not your fault, we have all go so used to Microsoft producing absolute rubbish and not clearly explaining anything, that we all think this is a good thing... We ALL need to think better and start to do things properly don't you agree?

now the link if you please?
dobdo99
 
Posts: 4
Joined: Mon Mar 05, 2012 7:51 pm

Re: Resul Expression of the Time Series Prediction with GP

Postby andrew.kirillov » Tue Mar 06, 2012 8:18 am

dobdo99 wrote:now the link if you please?

The link.
With best regards,
Andrew


Interested in supporting AForge.NET Framework?
User avatar
andrew.kirillov
Site Admin, AForge.NET Developer
 
Posts: 3446
Joined: Fri Jan 23, 2009 9:12 am
Location: UK

Re: Resul Expression of the Time Series Prediction with GP

Postby dobdo99 » Tue Mar 06, 2012 9:32 am

andrew.kirillov wrote:
dobdo99 wrote:code: returns solution of: $6 $0 +


The problem is that you don't try reading documentation and sample code. The documentation clearly says that $0 is x[t-1]. Now, if you have window size 5, then starting from $5 you have constants, which means that $6 is a second constant. Taking a very quick look at provided sample application, you can see that second constant is 2. So your expression is "x[t-1] + 2". This means that every new value of time series equals to previous one increased by 2. I think this statement is clear from the data sample you gave.

Here is the sample application code snipit

static void Main( )
{
double[] data = new double[13] { 1, 2, 4, 7, 11, 16, 22, 29, 37, 46, 56, 67, 79 };
// constants
double[] constants = new double[10] { 1, 2, 3, 5, 7, 11, 13, 17, 19, 23 };
// sliding window size
int windowSize = 5;
// create population

The second constant is 2, TRUE, what are these constant values? why do we have these specific numbers as constants?

why not double[] constants = new double[10] { 101, 102, 103, 105, 107, 1011, 1013, 1017, 1019, 1023 };

What is the relationship between constants and the time series functions?
dobdo99
 
Posts: 4
Joined: Mon Mar 05, 2012 7:51 pm



Next

Return to Artificial Intelligence