Hello again,

just for people who have a similar problem like me:

I think I'll create 2 subclasses ob GPTree, one for the non-based 
rules and one for the rulebased-rules. Using this method I can 
separate the evolution_state for the rules (I hope so).

I think I'll have to write a programm which creates the rules
at the end, but the information containing it, will be 
created by ECJ. I think I have to start the writer before 
each fitness test or I'll program a plugin between ecj and the 
rule engine.

Thanks to everybode who read this. If you got more questions 
to my project feel free to contact me.

Bojan

----- Ursprüngliche Mail -----
Von: "Bojan Janisch" <[log in to unmask]>
An: "ECJ Evolutionary Computation Toolkit" <[log in to unmask]>
Gesendet: Dienstag, 15. Januar 2013 15:33:09
Betreff: Re: Hierarchical GP

Not really. I just want to create rules. I'll set some 
rule types later. And in between I've read half of the 
manual, beginning to understand that it is possible 
with ecj. 

The ruleengine will be used to annotate clinical text. 
And I want to write my bachelor thesis about this 
project. I don't like the idea to reinvent the wheel, 
so I try something new. 

Back to your question:
I like the idea of Lenat also but I would not want 
to publish it, because I don't think that humans
are civilized enough to use these programs correctly.

Bojan

----- Ursprüngliche Mail -----
Von: "Mike Eggleston" <[log in to unmask]>
An: [log in to unmask]
Gesendet: Dienstag, 15. Januar 2013 15:04:34
Betreff: Re: Hierarchical GP

On Tue, 15 Jan 2013, Bojan Janisch might have said:

> Hello everyone,
> 
> my problem is that I want to use a hierarchical GP engine. 
> So is it possible with ecj to solve a GP-problem in a 
> GP-problem?
> 
> Background:
> I want to combine a rulebased method with GP to create and 
> extend rules automatically. The Problem is that I'm using 
> hierarchical rules, means rules that are based on other rules.
> 
> To solve this I must first calculate the non-based rules to a 
> best fit and than calculate the combination of these rules 
> to find the best rulebased rules.
> 
> So is ECJ capable of this?
> 
> In worst case I maybe have to write my own GP engine, that is 
> using ECJ for non-based rules, to find the best solution in 
> rulebased-rules.
> 
> Greets from Germany
> Bojan

Bojan,

Are you wanting to do something like Lenat's Eurisko?
I find Eurisko's idea very interesting.

Mike