Regarding Simplenlg use

[HEMANT POHARKAR]

Hello,

I came across one module while working on project in which I need to Generate text from Ontology(Owl file) and I might be having input in triples form or maybe in form of tuples(entities of ontology) . Will simplenlg would be compatible for implementing this module? Because I am not sure if we could use it for text generation from Owl file.

Can anyone help with issue? or suggest me some resource or reference for implementing my module?

Thanks,

Hemant

[Greg Wickham]

SimpleNLG would be fine for doing the surface realization, but by itself SimpleNLG does NOT deal with ontology and semantics.  You need another part to complete this pipeline.  I'm working on another system called Flex that uses WordNet and spreading activation to achieve this, but that part of Flex is not finished yet.

--
You received this message because you are subscribed to the Google Groups "simplenlg" group.
To unsubscribe from this group and stop receiving emails from it, send an email to simplenlg+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/simplenlg/b5bd7c05-8e73-4d5f-8741-b2d90027d74fn%40googlegroups.com.

[Saad Mahamood]

Hi Hemant,

Yes it is very much possible to use SimpleNLG to generate texts from an OWL file. Many academic and even commercial systems have incorporated SimpleNLG for that purpose. However SimpleNLG is just one component of a three stage process to generate a text:

1. Document Planning — Selection of content that you want to express in a document taken from your OWL file. Commonly structured as a graph tree with the leaf nodes representing individual sentences (e.g. Document->Section->Paragraph->[Sentence->OWL instance, Sentence->OWL instance, etc.].

2. Microplanning — For each given sentence the mapping of SimpleNLG syntax structures (Either expressed as XML or Java code). The idea is that for a given OWL instance your Microplanner would map the relevant SimpleNLG sentence syntax based on some conditional/rule logic (e.g. Map sentence structure X IF instance is Banana class). You can also pass values from the instance to populate values into the sentence syntax structure. The result from this is a syntactic tree input for the next step. It’s best to think of Microplanning as the conversion between the semantic Document Plan tree to SimpleNLG syntactic tree. 

3. Surface Realisation — SimpleNLG deals with this and uses the syntactic tree input to generate text output.

Hopefully, this explanation will make a little more sense and give you a bit guidance on how to build your system.

Many Thanks,

Saad Mahamood.

[HEMANT POHARKAR]

Thank you so much Sir for the guidance. It really helped me a lot.