Quantcast

Isomorphic taxonomies

classic Classic list List threaded Threaded
15 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Isomorphic taxonomies

Marcelino Borges
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Jim Balhoff-2
Hi,

We deal with something similar by generating taxonomies automatically as much as possible, using a template system. For example, an taxonomy of environmental exposures can be driven by taxonomies of various stressors (e.g. chemicals). Given an “exposure” pattern/model, a list of chemical fillers, and a chemical ontology, one can infer the chemical exposure taxonomy. Here is an ontology being developed this way:


The chemical exposures are managed via a simple tab-delimited file which references chemical terms:


A pattern template is used to generate the chemical exposure concepts:


This is done via this system, using a build tool:


Best regards,
Jim



On Apr 19, 2017, at 2:39 PM, Marcelino Borges <[hidden email]> wrote:

Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Alex Shkotin
In reply to this post by Marcelino Borges
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Marcelino Borges
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Alex Shkotin
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Marcelino Borges
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Alex Shkotin
At least you may have definitions for your classes with anonymous individuals coupled with pictures like this:
DogImage(x) == some x: Dog(x) and on(x pic1) 
But I doubt it's a good idea to have class DogImage

Alex


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

samsontu
You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Chris Mungall-2

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Marcelino Borges
Yes, Chris and Samson.
It is possible to define a mirror taxonomy in many ways and including it in the ontology.
But I wonder if having all these isomorphic taxonomies is the best way of modeling this. Maybe there is some better way of solving this issue, keeping only the target taxonomy and using some other aditional design patterns for representing the same knoeledge. Maybe there is some ontology design pattern for solving this recurrent issue, without replicating the taxonomy.

Best regards.





2017-04-24 21:44 GMT-03:00 Chris Mungall <[hidden email]>:

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Csongor Nyulas
Administrator
Marcelino,
I would like to point out to you that is really hard to give you suggestions for possible solutions to a problem that you did not specify.
You say "But I wonder if having all these isomorphic taxonomies is the best way of modeling this."
What is the "this"?
What is your end goal? What are you trying to achieve, overall, using such isomorphic taxonomies?

Csongor


On 04/25/2017 07:46 PM, Marcelino Borges wrote:
Yes, Chris and Samson.
It is possible to define a mirror taxonomy in many ways and including it in the ontology.
But I wonder if having all these isomorphic taxonomies is the best way of modeling this. Maybe there is some better way of solving this issue, keeping only the target taxonomy and using some other aditional design patterns for representing the same knoeledge. Maybe there is some ontology design pattern for solving this recurrent issue, without replicating the taxonomy.

Best regards.





2017-04-24 21:44 GMT-03:00 Chris Mungall <[hidden email]>:

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user




_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Biswanath Dutta


Hi Marcelino, 
    In continuation to Csongor's remarks, I would also like to understand the need of maintaining the two taxonomies (one for the animals and the other one for the animal images). What is the task? What is the application goal? 
I think if we have clear information on these, we would be able to understand the problem in a better way. And probably this will enable us to design a better model. 

With regards,
Biswanath Dutta 



On Wed, Apr 26, 2017 at 10:02 AM, Csongor Nyulas <[hidden email]> wrote:
Marcelino,
I would like to point out to you that is really hard to give you suggestions for possible solutions to a problem that you did not specify.
You say "But I wonder if having all these isomorphic taxonomies is the best way of modeling this."
What is the "this"?
What is your end goal? What are you trying to achieve, overall, using such isomorphic taxonomies?

Csongor



On 04/25/2017 07:46 PM, Marcelino Borges wrote:
Yes, Chris and Samson.
It is possible to define a mirror taxonomy in many ways and including it in the ontology.
But I wonder if having all these isomorphic taxonomies is the best way of modeling this. Maybe there is some better way of solving this issue, keeping only the target taxonomy and using some other aditional design patterns for representing the same knoeledge. Maybe there is some ontology design pattern for solving this recurrent issue, without replicating the taxonomy.

Best regards.





2017-04-24 21:44 GMT-03:00 Chris Mungall <[hidden email]>:

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user




_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Marcelino Borges
It is an application for the domain of Biology.
I need a taxonomy of animals because I need to store and reason about individual animals.
And I need a taxonomy of animal images because I need to store and reason about animal images. Some animal have animal imagens associated. But I also have some images that are not related to instances of animals. And I would like to be able to perform queries like "give me images of mammals", where the resulting images include images of dogs, cats, cows, etc.

Best regards.

2017-04-26 2:16 GMT-03:00 Biswanath Dutta <[hidden email]>:


Hi Marcelino, 
    In continuation to Csongor's remarks, I would also like to understand the need of maintaining the two taxonomies (one for the animals and the other one for the animal images). What is the task? What is the application goal? 
I think if we have clear information on these, we would be able to understand the problem in a better way. And probably this will enable us to design a better model. 

With regards,
Biswanath Dutta 



On Wed, Apr 26, 2017 at 10:02 AM, Csongor Nyulas <[hidden email]> wrote:
Marcelino,
I would like to point out to you that is really hard to give you suggestions for possible solutions to a problem that you did not specify.
You say "But I wonder if having all these isomorphic taxonomies is the best way of modeling this."
What is the "this"?
What is your end goal? What are you trying to achieve, overall, using such isomorphic taxonomies?

Csongor



On 04/25/2017 07:46 PM, Marcelino Borges wrote:
Yes, Chris and Samson.
It is possible to define a mirror taxonomy in many ways and including it in the ontology.
But I wonder if having all these isomorphic taxonomies is the best way of modeling this. Maybe there is some better way of solving this issue, keeping only the target taxonomy and using some other aditional design patterns for representing the same knoeledge. Maybe there is some ontology design pattern for solving this recurrent issue, without replicating the taxonomy.

Best regards.





2017-04-24 21:44 GMT-03:00 Chris Mungall <[hidden email]>:

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user




_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

samsontu
For making queries of the kind you mentioned, the better solution may depends on your query language. You can make your images individuals of the type (Image and (is_image_of some X)), where X is a class from your Animal (and other things that have images) hierarchy. You can easily make DL queries for instances of (Image and (is_image_of some X)). You don’t have to maintain a parallel hierarchy in this case. However, if you have to use SPARQL, your query expression will be much more complex, as was discussed in another thread on "Can Sparql query information about object property restriction.” That would be a reason for maintaining parallel named hierarchies.

With best regards,
Samson



On Apr 26, 2017, at 6:26 AM, Marcelino Borges <[hidden email]> wrote:

It is an application for the domain of Biology.
I need a taxonomy of animals because I need to store and reason about individual animals.
And I need a taxonomy of animal images because I need to store and reason about animal images. Some animal have animal imagens associated. But I also have some images that are not related to instances of animals. And I would like to be able to perform queries like "give me images of mammals", where the resulting images include images of dogs, cats, cows, etc.

Best regards.

2017-04-26 2:16 GMT-03:00 Biswanath Dutta <[hidden email]>:


Hi Marcelino, 
    In continuation to Csongor's remarks, I would also like to understand the need of maintaining the two taxonomies (one for the animals and the other one for the animal images). What is the task? What is the application goal? 
I think if we have clear information on these, we would be able to understand the problem in a better way. And probably this will enable us to design a better model. 

With regards,
Biswanath Dutta 



On Wed, Apr 26, 2017 at 10:02 AM, Csongor Nyulas <[hidden email]> wrote:
Marcelino,
I would like to point out to you that is really hard to give you suggestions for possible solutions to a problem that you did not specify.
You say "But I wonder if having all these isomorphic taxonomies is the best way of modeling this."
What is the "this"?
What is your end goal? What are you trying to achieve, overall, using such isomorphic taxonomies?

Csongor



On 04/25/2017 07:46 PM, Marcelino Borges wrote:
Yes, Chris and Samson.
It is possible to define a mirror taxonomy in many ways and including it in the ontology.
But I wonder if having all these isomorphic taxonomies is the best way of modeling this. Maybe there is some better way of solving this issue, keeping only the target taxonomy and using some other aditional design patterns for representing the same knoeledge. Maybe there is some ontology design pattern for solving this recurrent issue, without replicating the taxonomy.

Best regards.





2017-04-24 21:44 GMT-03:00 Chris Mungall <[hidden email]>:

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user




_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

-- 
Samson Tu                                                      email: [hidden email]
Senior Research Engineer                              web: www.stanford.edu/~swt/
Center for Biomedical Informatics Research  phone: 1-650-725-3391
Stanford University                                          fax: 1-650-725-7944




_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Isomorphic taxonomies

Biswanath Dutta
In reply to this post by Marcelino Borges


Dear Marcelino, 
          IMO, You can do the followings: 
(1) Create a taxonomy for Animals. 
(2) Create a class "Image" (alternatively, you can even create a taxonomy for "Images.". BUT this taxonomy should not be built by Animal types, rather try to classify the images from different perspectives, form instance, by quality). 
(3) Create an object property, say "isImageOf" with "domain:Image" and "range:Animal". (you can apply "some" restriction) 

I guess this helps. 

With regards,
Biswanath Dutta 



On Wed, Apr 26, 2017 at 6:56 PM, Marcelino Borges <[hidden email]> wrote:
It is an application for the domain of Biology.
I need a taxonomy of animals because I need to store and reason about individual animals.
And I need a taxonomy of animal images because I need to store and reason about animal images. Some animal have animal imagens associated. But I also have some images that are not related to instances of animals. And I would like to be able to perform queries like "give me images of mammals", where the resulting images include images of dogs, cats, cows, etc.

Best regards.

2017-04-26 2:16 GMT-03:00 Biswanath Dutta <[hidden email]>:


Hi Marcelino, 
    In continuation to Csongor's remarks, I would also like to understand the need of maintaining the two taxonomies (one for the animals and the other one for the animal images). What is the task? What is the application goal? 
I think if we have clear information on these, we would be able to understand the problem in a better way. And probably this will enable us to design a better model. 

With regards,
Biswanath Dutta 



On Wed, Apr 26, 2017 at 10:02 AM, Csongor Nyulas <[hidden email]> wrote:
Marcelino,
I would like to point out to you that is really hard to give you suggestions for possible solutions to a problem that you did not specify.
You say "But I wonder if having all these isomorphic taxonomies is the best way of modeling this."
What is the "this"?
What is your end goal? What are you trying to achieve, overall, using such isomorphic taxonomies?

Csongor



On 04/25/2017 07:46 PM, Marcelino Borges wrote:
Yes, Chris and Samson.
It is possible to define a mirror taxonomy in many ways and including it in the ontology.
But I wonder if having all these isomorphic taxonomies is the best way of modeling this. Maybe there is some better way of solving this issue, keeping only the target taxonomy and using some other aditional design patterns for representing the same knoeledge. Maybe there is some ontology design pattern for solving this recurrent issue, without replicating the taxonomy.

Best regards.





2017-04-24 21:44 GMT-03:00 Chris Mungall <[hidden email]>:

Yes, see also Jim Balhoff's response in this thread. Using the toolkit he mentions doesn't require any programming.

You would just have a yaml config:

name:
  text: "%s image"
  vars:
    - animal

equivalentTo: 
  text: "'image' and 'depicts' some %s"
  vars:
    - animal

And feed in your list of animals - instant shadow hierarchy!

On 24 Apr 2017, at 15:12, Samson Tu wrote:

You can create manually or programmatically classes Ximage (AnimalImage, MammalImage, DogImmage etc.) equivalent to (Image and (is-image-of some X). The reasoner will maintain the image taxonomy in parallel with the animal taxonomy.

With best regards,
Samson


2017-04-24 19:41 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

OK. I can import an animal taxonomy.
However, I need the two taxonomies in my ontology: an animal taxonomy and a animal image taxonomy. Is there any way of keeping the two taxonomies without replicating them?

Let us consider that I have the following taxonomy:
-Animal
--Mammal
---Dog

I also need the following animal image taxonomy:
-AnimalImagem
--MammalImge
---DogImage

The straightforward way of doing this is developing the two taxonomies, as independent ones. However, I was trying to find some way of taking advantage of the structure of the animal taxonomy for developing the animal image taxonomy. Is there any smarter way of doing this?

Best regards.

2017-04-24 13:33 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

I did not get that you need to develop an animal ontology. Usually, it's better to import existing one.

Best regards,

Alex

2017-04-24 17:47 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi Alex.

I'm not sure if I understood.

I need to develop a single ontology that includes both taxonomies: animal taxonomy and animal image taxonomy. It is not clear for me how can I deal with this using imports. It is important to notice that I need to distinguish instances of animals and instances of animal images. And there are some instances of animal image that are not related to instances of animals, in the knowledge base.

Best regards.



2017-04-21 10:05 GMT-03:00 Alex Shkotin <[hidden email]>:
Hi Marcelino,

It's better to think that we have animal taxonomy gratis for our images. In OWL2 there is an import possibility for this case. 
So we have animal class, images class, and object relation like _on_.
We can import animal taxonomy and have image taxonomy gratis.

Alex

2017-04-19 21:39 GMT+03:00 Marcelino Borges <[hidden email]>:
Hi all.

In several different contexts I have found situations in which we have two or more taxonomies of concepts that are isomorphic, and seem to replicate a taxonomy, with some semantic variations.

For example, we can imagine a taxonomy about animals (that will have concepts like cat, dog, snake, etc), and we can imagine also a taxonomy about images of animals (that will have concepts like, cat image, dog image, etc). 

An intuitive way of dealing with this is to have two separate taxonomies, since the instances of animals are not instances of animal imagens (of course). But I'm not happy with this kind of modeling. It seems that there is some smarter way of doing this.

Other similar example is a scenario where we need to deal with a taxonomy of actual objects and a taxonomy of conceptualized objects. There are other similar scenarios.

There are smarter ways of dealing with this?

Best regards.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user




_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Loading...