Modeling substances

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

Modeling substances

Marcelino Borges
Hi.

What are the possibilities of modeling substances (rock, wood, water, gold, etc) in OWL ontologies, and what are the advantages and disadvantages of each alternative?

I think that what we need to decide in this case is what are the instances of classes representing substances.

And, if we don't have instances for these classes, maybe they don't have to be modeled as classes. But, in this case, how to model them? And, in some cases, we have a complex hierarchy and attributes that describe the different sub-types. For example, in the case of rocks. Geologists have a complex hierarchy for rocks and several attributes for describing rocks.

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: Modeling substances

Csongor Nyulas
Administrator
Modeling substances is not a simple problem, indeed, but without going into deep analysis of the modeling alternatives, I just say this quickly:

If you want to express the fact that these substances can have subtypes, you should represent them as classes. Even if you are not interested in modeling any particular instance of those classes, they do in fact have instances (the imaginary piece of rock on my desk is an instance of the your Rock class). You can then describe these classes with the properties that they have, and you can reason about them.

Csongor

On 03/16/2017 12:09 PM, Marcelino Borges wrote:
Hi.

What are the possibilities of modeling substances (rock, wood, water, gold, etc) in OWL ontologies, and what are the advantages and disadvantages of each alternative?

I think that what we need to decide in this case is what are the instances of classes representing substances.

And, if we don't have instances for these classes, maybe they don't have to be modeled as classes. But, in this case, how to model them? And, in some cases, we have a complex hierarchy and attributes that describe the different sub-types. For example, in the case of rocks. Geologists have a complex hierarchy for rocks and several attributes for describing rocks.

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: Modeling substances

Ramona Walls
Marcelino,

You should have a look at the Environment Ontology and see how they have modeled substances for the purposes of describing where organisms live: http://environmentontology.org/

Ramona

------------------------------------------------------
Ramona L. Walls, Ph.D.
Senior Scientific Analyst, CyVerse, University of Arizona
Research Associate, Bio5 Institute, University of Arizona

On Thu, Mar 16, 2017 at 12:44 PM, Csongor Nyulas <[hidden email]> wrote:
Modeling substances is not a simple problem, indeed, but without going into deep analysis of the modeling alternatives, I just say this quickly:

If you want to express the fact that these substances can have subtypes, you should represent them as classes. Even if you are not interested in modeling any particular instance of those classes, they do in fact have instances (the imaginary piece of rock on my desk is an instance of the your Rock class). You can then describe these classes with the properties that they have, and you can reason about them.

Csongor


On 03/16/2017 12:09 PM, Marcelino Borges wrote:
Hi.

What are the possibilities of modeling substances (rock, wood, water, gold, etc) in OWL ontologies, and what are the advantages and disadvantages of each alternative?

I think that what we need to decide in this case is what are the instances of classes representing substances.

And, if we don't have instances for these classes, maybe they don't have to be modeled as classes. But, in this case, how to model them? And, in some cases, we have a complex hierarchy and attributes that describe the different sub-types. For example, in the case of rocks. Geologists have a complex hierarchy for rocks and several attributes for describing rocks.

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: Modeling substances

Alex Shkotin
Hi Marcelino!

As you mentioned rocks, have a look at our study of rock type
definitions formalization
https://www.researchgate.net/publication/51910491_Towards_OWL-based_Knowledge_Representation_in_Petrology

Alex

2017-03-16 23:57 GMT+04:00, Ramona Walls <[hidden email]>:

> Marcelino,
>
> You should have a look at the Environment Ontology and see how they have
> modeled substances for the purposes of describing where organisms live:
> http://environmentontology.org/
>
> Ramona
>
> ------------------------------------------------------
> Ramona L. Walls, Ph.D.
> Senior Scientific Analyst, CyVerse, University of Arizona
> Research Associate, Bio5 Institute, University of Arizona
>
> On Thu, Mar 16, 2017 at 12:44 PM, Csongor Nyulas <
> [hidden email]> wrote:
>
>> Modeling substances is not a simple problem, indeed, but without going
>> into deep analysis of the modeling alternatives, I just say this quickly:
>>
>> If you want to express the fact that these substances can have subtypes,
>> you should represent them as classes. Even if you are not interested in
>> modeling any particular instance of those classes, they do in fact have
>> instances (the imaginary piece of rock on my desk is an instance of the
>> your Rock class). You can then describe these classes with the properties
>> that they have, and you can reason about them.
>>
>> Csongor
>>
>>
>> On 03/16/2017 12:09 PM, Marcelino Borges wrote:
>>
>> Hi.
>>
>> What are the possibilities of modeling substances (rock, wood, water,
>> gold, etc) in OWL ontologies, and what are the advantages and
>> disadvantages
>> of each alternative?
>>
>> I think that what we need to decide in this case is what are the
>> instances
>> of classes representing substances.
>>
>> And, if we don't have instances for these classes, maybe they don't have
>> to be modeled as classes. But, in this case, how to model them? And, in
>> some cases, we have a complex hierarchy and attributes that describe the
>> different sub-types. For example, in the case of rocks. Geologists have a
>> complex hierarchy for rocks and several attributes for describing rocks.
>>
>> Best regards.
>>
>>
>> _______________________________________________
>> protege-user mailing
>> [hidden email]://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: Modeling substances

Michael Osborne
You mention Geology as a use case, but you may be able to learn from the medical domain. SNOMED CT has a substance hierarchy - it's mainly there to support the modeling of Clinical useful classes, however it may have some of the content you need.
There is a project run by SNOMED International to restructure the hierarchy (it currently suffers from inconsistently modeled content)
and
you can of course browse the Substance hierarchy here:
or here for a graph view:
Cheers
Michael

On 18 March 2017 at 03:32, Alex Shkotin <[hidden email]> wrote:
Hi Marcelino!

As you mentioned rocks, have a look at our study of rock type
definitions formalization
https://www.researchgate.net/publication/51910491_Towards_OWL-based_Knowledge_Representation_in_Petrology

Alex

2017-03-16 23:57 GMT+04:00, Ramona Walls <[hidden email]>:
> Marcelino,
>
> You should have a look at the Environment Ontology and see how they have
> modeled substances for the purposes of describing where organisms live:
> http://environmentontology.org/
>
> Ramona
>
> ------------------------------------------------------
> Ramona L. Walls, Ph.D.
> Senior Scientific Analyst, CyVerse, University of Arizona
> Research Associate, Bio5 Institute, University of Arizona
>
> On Thu, Mar 16, 2017 at 12:44 PM, Csongor Nyulas <
> [hidden email]> wrote:
>
>> Modeling substances is not a simple problem, indeed, but without going
>> into deep analysis of the modeling alternatives, I just say this quickly:
>>
>> If you want to express the fact that these substances can have subtypes,
>> you should represent them as classes. Even if you are not interested in
>> modeling any particular instance of those classes, they do in fact have
>> instances (the imaginary piece of rock on my desk is an instance of the
>> your Rock class). You can then describe these classes with the properties
>> that they have, and you can reason about them.
>>
>> Csongor
>>
>>
>> On 03/16/2017 12:09 PM, Marcelino Borges wrote:
>>
>> Hi.
>>
>> What are the possibilities of modeling substances (rock, wood, water,
>> gold, etc) in OWL ontologies, and what are the advantages and
>> disadvantages
>> of each alternative?
>>
>> I think that what we need to decide in this case is what are the
>> instances
>> of classes representing substances.
>>
>> And, if we don't have instances for these classes, maybe they don't have
>> to be modeled as classes. But, in this case, how to model them? And, in
>> some cases, we have a complex hierarchy and attributes that describe the
>> different sub-types. For example, in the case of rocks. Geologists have a
>> complex hierarchy for rocks and several attributes for describing rocks.
>>
>> Best regards.
>>
>>
>> _______________________________________________
>> protege-user mailing
>> listprotege-user@lists.stanford.eduhttps://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



--
Michael Osborne

_______________________________________________
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: Modeling substances

Marcelino Borges
Thank you for the contributions.

I agree that substances have instances, if we interpret these classes as "Portion of" (or "Quantity of", as said here).

However, for example, it is not clear how to model, for example, that a given bicycle is made of Steel, Plastic, Rubber, etc. I don't know how to model this (in OWL) if I consider just the Substance classes, without instances. 

On the other hand, if I consider the instances, it is hard to think about how many instances of each material are considered in a given situation. For example, let us imagine the instance b1 of Bicycle. And let us consider that we need to say that b1 is made of Steel (and other substances). But how many instances of Steel are related to b1? Just one instance that is the mereological sum of every portion of Steel present in b1?

Best regards.

2017-03-17 20:58 GMT-03:00 Michael Osborne <[hidden email]>:
You mention Geology as a use case, but you may be able to learn from the medical domain. SNOMED CT has a substance hierarchy - it's mainly there to support the modeling of Clinical useful classes, however it may have some of the content you need.
There is a project run by SNOMED International to restructure the hierarchy (it currently suffers from inconsistently modeled content)
and
you can of course browse the Substance hierarchy here:
or here for a graph view:
Cheers
Michael

On 18 March 2017 at 03:32, Alex Shkotin <[hidden email]> wrote:
Hi Marcelino!

As you mentioned rocks, have a look at our study of rock type
definitions formalization
https://www.researchgate.net/publication/51910491_Towards_OWL-based_Knowledge_Representation_in_Petrology

Alex

2017-03-16 23:57 GMT+04:00, Ramona Walls <[hidden email]>:
> Marcelino,
>
> You should have a look at the Environment Ontology and see how they have
> modeled substances for the purposes of describing where organisms live:
> http://environmentontology.org/
>
> Ramona
>
> ------------------------------------------------------
> Ramona L. Walls, Ph.D.
> Senior Scientific Analyst, CyVerse, University of Arizona
> Research Associate, Bio5 Institute, University of Arizona
>
> On Thu, Mar 16, 2017 at 12:44 PM, Csongor Nyulas <
> [hidden email]> wrote:
>
>> Modeling substances is not a simple problem, indeed, but without going
>> into deep analysis of the modeling alternatives, I just say this quickly:
>>
>> If you want to express the fact that these substances can have subtypes,
>> you should represent them as classes. Even if you are not interested in
>> modeling any particular instance of those classes, they do in fact have
>> instances (the imaginary piece of rock on my desk is an instance of the
>> your Rock class). You can then describe these classes with the properties
>> that they have, and you can reason about them.
>>
>> Csongor
>>
>>
>> On 03/16/2017 12:09 PM, Marcelino Borges wrote:
>>
>> Hi.
>>
>> What are the possibilities of modeling substances (rock, wood, water,
>> gold, etc) in OWL ontologies, and what are the advantages and
>> disadvantages
>> of each alternative?
>>
>> I think that what we need to decide in this case is what are the
>> instances
>> of classes representing substances.
>>
>> And, if we don't have instances for these classes, maybe they don't have
>> to be modeled as classes. But, in this case, how to model them? And, in
>> some cases, we have a complex hierarchy and attributes that describe the
>> different sub-types. For example, in the case of rocks. Geologists have a
>> complex hierarchy for rocks and several attributes for describing rocks.
>>
>> Best regards.
>>
>>
>> _______________________________________________
>> protege-user mailing
>> [hidden email]d.eduhttps://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



--
Michael Osborne

_______________________________________________
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: Modeling substances

Csongor Nyulas
Administrator
The nice thing about OWL, is that you can describe things at class level, without having to have instances in your ontology.
The SomeValuesFrom class expression, represented in Manchester syntax with the keyword "some", allows you to talk about some undefined set of instances of a certain class. For example you could have the following class hierarchy
Bycicle
    CrossBycicle
    ElecrtricBycicle
    HybridBycicle
    MountainBike
    ...
    RoadBike
        Trek_RoadBike
            TREK_MADONE_9.5_ULTEGRA_DI2-2017
            TREK_PROCALIBER_9.9_SL_RACE_SHOP_LIMITED-2017
            ...
Material
    Steel
    Glass
    GlassFiber
    CarbonFiber
        OCLV_CarbonFiber
            600_Series_OCLV_Carbon
            ...
        TCT_CarbonFiber
        Madone_KVF_Full_Carbon
        ...
    ...


Then you can define the class TREK_MADONE_9.5_ULTEGRA_DI2-2017 with axioms such:
    TREK_MADONE_9.5_ULTEGRA_DI2-2017 SubClassOf (hasMaterial some 600_Series_OCLV_Carbon) and
    TREK_MADONE_9.5_ULTEGRA_DI2-2017 SubClassOf (hasMaterial some Madone_KVF_Full_Carbon) and
    ...


This means that any instance of the class TREK_MADONE_9.5_ULTEGRA_DI2-2017 sitting in someone's garage will have some instances of 600_Series_OCLV_Carbon material in it. You don't need to have those instances explicitly represented in your ontology. If you happen to have an individual marcelinos_trek_madone1 in your ontology, which is an instance of the class TREK_MADONE_9.5_ULTEGRA_DI2-2017, when you query for all the things that have material 600_Series_OCLV_Carbonmarcelinos_trek_madone1 will be returned. And marcelinos_trek_madone1 will be returned also if you query for all things that have CarbonFiber as their material (you could use this query in the DL Query tab "hasMaterial some CarbonFiber", and select the "Instances" check box).

Csongor


On 03/21/2017 09:31 AM, Marcelino Borges wrote:
Thank you for the contributions.

I agree that substances have instances, if we interpret these classes as "Portion of" (or "Quantity of", as said here).

However, for example, it is not clear how to model, for example, that a given bicycle is made of Steel, Plastic, Rubber, etc. I don't know how to model this (in OWL) if I consider just the Substance classes, without instances. 

On the other hand, if I consider the instances, it is hard to think about how many instances of each material are considered in a given situation. For example, let us imagine the instance b1 of Bicycle. And let us consider that we need to say that b1 is made of Steel (and other substances). But how many instances of Steel are related to b1? Just one instance that is the mereological sum of every portion of Steel present in b1?

Best regards.

2017-03-17 20:58 GMT-03:00 Michael Osborne <[hidden email]>:
You mention Geology as a use case, but you may be able to learn from the medical domain. SNOMED CT has a substance hierarchy - it's mainly there to support the modeling of Clinical useful classes, however it may have some of the content you need.
There is a project run by SNOMED International to restructure the hierarchy (it currently suffers from inconsistently modeled content)
and
you can of course browse the Substance hierarchy here:
or here for a graph view:
Cheers
Michael

On 18 March 2017 at 03:32, Alex Shkotin <[hidden email]> wrote:
Hi Marcelino!

As you mentioned rocks, have a look at our study of rock type
definitions formalization
https://www.researchgate.net/publication/51910491_Towards_OWL-based_Knowledge_Representation_in_Petrology

Alex

2017-03-16 23:57 GMT+04:00, Ramona Walls <[hidden email]>:
> Marcelino,
>
> You should have a look at the Environment Ontology and see how they have
> modeled substances for the purposes of describing where organisms live:
> http://environmentontology.org/
>
> Ramona
>
> ------------------------------------------------------
> Ramona L. Walls, Ph.D.
> Senior Scientific Analyst, CyVerse, University of Arizona
> Research Associate, Bio5 Institute, University of Arizona
>
> On Thu, Mar 16, 2017 at 12:44 PM, Csongor Nyulas <
> [hidden email]> wrote:
>
>> Modeling substances is not a simple problem, indeed, but without going
>> into deep analysis of the modeling alternatives, I just say this quickly:
>>
>> If you want to express the fact that these substances can have subtypes,
>> you should represent them as classes. Even if you are not interested in
>> modeling any particular instance of those classes, they do in fact have
>> instances (the imaginary piece of rock on my desk is an instance of the
>> your Rock class). You can then describe these classes with the properties
>> that they have, and you can reason about them.
>>
>> Csongor
>>
>>
>> On 03/16/2017 12:09 PM, Marcelino Borges wrote:
>>
>> Hi.
>>
>> What are the possibilities of modeling substances (rock, wood, water,
>> gold, etc) in OWL ontologies, and what are the advantages and
>> disadvantages
>> of each alternative?
>>
>> I think that what we need to decide in this case is what are the
>> instances
>> of classes representing substances.
>>
>> And, if we don't have instances for these classes, maybe they don't have
>> to be modeled as classes. But, in this case, how to model them? And, in
>> some cases, we have a complex hierarchy and attributes that describe the
>> different sub-types. For example, in the case of rocks. Geologists have a
>> complex hierarchy for rocks and several attributes for describing rocks.
>>
>> Best regards.
>>
>>
>> _______________________________________________
>> protege-user mailing
>> [hidden email]d.eduhttps://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



--
Michael Osborne

_______________________________________________
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: Modeling substances

Marcelino Borges
Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.

2017-03-21 15:40 GMT-03:00 Csongor Nyulas <[hidden email]>:
The nice thing about OWL, is that you can describe things at class level, without having to have instances in your ontology.
The SomeValuesFrom class expression, represented in Manchester syntax with the keyword "some", allows you to talk about some undefined set of instances of a certain class. For example you could have the following class hierarchy
Bycicle
    CrossBycicle
    ElecrtricBycicle
    HybridBycicle
    MountainBike
    ...
    RoadBike
        Trek_RoadBike
            TREK_MADONE_9.5_ULTEGRA_DI2-2017
            TREK_PROCALIBER_9.9_SL_RACE_SHOP_LIMITED-2017
            ...
Material
    Steel
    Glass
    GlassFiber
    CarbonFiber
        OCLV_CarbonFiber
            600_Series_OCLV_Carbon
            ...
        TCT_CarbonFiber
        Madone_KVF_Full_Carbon
        ...
    ...


Then you can define the class TREK_MADONE_9.5_ULTEGRA_DI2-2017 with axioms such:
    TREK_MADONE_9.5_ULTEGRA_DI2-2017 SubClassOf (hasMaterial some 600_Series_OCLV_Carbon) and
    TREK_MADONE_9.5_ULTEGRA_DI2-2017 SubClassOf (hasMaterial some Madone_KVF_Full_Carbon) and
    ...


This means that any instance of the class TREK_MADONE_9.5_ULTEGRA_DI2-2017 sitting in someone's garage will have some instances of 600_Series_OCLV_Carbon material in it. You don't need to have those instances explicitly represented in your ontology. If you happen to have an individual marcelinos_trek_madone1 in your ontology, which is an instance of the class TREK_MADONE_9.5_ULTEGRA_DI2-2017, when you query for all the things that have material 600_Series_OCLV_Carbonmarcelinos_trek_madone1 will be returned. And marcelinos_trek_madone1 will be returned also if you query for all things that have CarbonFiber as their material (you could use this query in the DL Query tab "hasMaterial some CarbonFiber", and select the "Instances" check box).

Csongor



On 03/21/2017 09:31 AM, Marcelino Borges wrote:
Thank you for the contributions.

I agree that substances have instances, if we interpret these classes as "Portion of" (or "Quantity of", as said here).

However, for example, it is not clear how to model, for example, that a given bicycle is made of Steel, Plastic, Rubber, etc. I don't know how to model this (in OWL) if I consider just the Substance classes, without instances. 

On the other hand, if I consider the instances, it is hard to think about how many instances of each material are considered in a given situation. For example, let us imagine the instance b1 of Bicycle. And let us consider that we need to say that b1 is made of Steel (and other substances). But how many instances of Steel are related to b1? Just one instance that is the mereological sum of every portion of Steel present in b1?

Best regards.

2017-03-17 20:58 GMT-03:00 Michael Osborne <[hidden email]>:
You mention Geology as a use case, but you may be able to learn from the medical domain. SNOMED CT has a substance hierarchy - it's mainly there to support the modeling of Clinical useful classes, however it may have some of the content you need.
There is a project run by SNOMED International to restructure the hierarchy (it currently suffers from inconsistently modeled content)
and
you can of course browse the Substance hierarchy here:
or here for a graph view:
Cheers
Michael

On 18 March 2017 at 03:32, Alex Shkotin <[hidden email]> wrote:
Hi Marcelino!

As you mentioned rocks, have a look at our study of rock type
definitions formalization
https://www.researchgate.net/publication/51910491_Towards_OWL-based_Knowledge_Representation_in_Petrology

Alex

2017-03-16 23:57 GMT+04:00, Ramona Walls <[hidden email]>:
> Marcelino,
>
> You should have a look at the Environment Ontology and see how they have
> modeled substances for the purposes of describing where organisms live:
> http://environmentontology.org/
>
> Ramona
>
> ------------------------------------------------------
> Ramona L. Walls, Ph.D.
> Senior Scientific Analyst, CyVerse, University of Arizona
> Research Associate, Bio5 Institute, University of Arizona
>
> On Thu, Mar 16, 2017 at 12:44 PM, Csongor Nyulas <
> [hidden email]> wrote:
>
>> Modeling substances is not a simple problem, indeed, but without going
>> into deep analysis of the modeling alternatives, I just say this quickly:
>>
>> If you want to express the fact that these substances can have subtypes,
>> you should represent them as classes. Even if you are not interested in
>> modeling any particular instance of those classes, they do in fact have
>> instances (the imaginary piece of rock on my desk is an instance of the
>> your Rock class). You can then describe these classes with the properties
>> that they have, and you can reason about them.
>>
>> Csongor
>>
>>
>> On 03/16/2017 12:09 PM, Marcelino Borges wrote:
>>
>> Hi.
>>
>> What are the possibilities of modeling substances (rock, wood, water,
>> gold, etc) in OWL ontologies, and what are the advantages and
>> disadvantages
>> of each alternative?
>>
>> I think that what we need to decide in this case is what are the
>> instances
>> of classes representing substances.
>>
>> And, if we don't have instances for these classes, maybe they don't have
>> to be modeled as classes. But, in this case, how to model them? And, in
>> some cases, we have a complex hierarchy and attributes that describe the
>> different sub-types. For example, in the case of rocks. Geologists have a
>> complex hierarchy for rocks and several attributes for describing rocks.
>>
>> Best regards.
>>
>>
>> _______________________________________________
>> protege-user mailing
>> [hidden email]d.eduhttps://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



--
Michael Osborne

_______________________________________________
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: Modeling substances

samsontu

On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:

Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.


Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)

Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.

There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.

With best regards,
Samson

_______________________________________________
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: Modeling substances

Marcelino Borges
Yes, Samson.
I agree with you. According to my intuition, this case can only be handled by considering individuals.

Best regards.

2017-03-22 13:52 GMT-03:00 Samson Tu <[hidden email]>:

On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:

Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.


Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)

Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.

There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.

With best regards,
Samson

_______________________________________________
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: Modeling substances

Csongor Nyulas
Administrator
This can be modelled in pure OWL with property chains as well. [1] [2]
For example you could say that:

    owns o madeOf subPropertyOf: owns


Csongor

[1] https://www.w3.org/TR/owl2-primer/#Property_Chains
[2] http://protegeproject.github.io/protege/views/object-property-description/


On 03/22/2017 12:00 PM, Marcelino Borges wrote:
Yes, Samson.
I agree with you. According to my intuition, this case can only be handled by considering individuals.

Best regards.

2017-03-22 13:52 GMT-03:00 Samson Tu <[hidden email]>:

On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:

Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.


Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)

Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.

There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.

With best regards,
Samson

_______________________________________________
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: Modeling substances

Marcelino Borges
Yes.
Indeed, I had in mind the option with property chains.
But the point is that, in this case, we need to deal with individuals and not only with classes.

2017-03-22 16:26 GMT-03:00 Csongor Nyulas <[hidden email]>:
This can be modelled in pure OWL with property chains as well. [1] [2]
For example you could say that:

    owns o madeOf subPropertyOf: owns


Csongor

[1] https://www.w3.org/TR/owl2-primer/#Property_Chains
[2] http://protegeproject.github.io/protege/views/object-property-description/



On 03/22/2017 12:00 PM, Marcelino Borges wrote:
Yes, Samson.
I agree with you. According to my intuition, this case can only be handled by considering individuals.

Best regards.

2017-03-22 13:52 GMT-03:00 Samson Tu <[hidden email]>:

On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:

Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.


Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)

Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.

There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.

With best regards,
Samson

_______________________________________________
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: Modeling substances

samsontu
In reply to this post by Marcelino Borges
Maybe I should be more explicit. Using OWL class expressions, your case can be handled without any instance of Steel. The SWRL rule involves individual Person and Bicycle, but not any Steel individual.

With best regards,
Samson

On Mar 22, 2017, at 12:00 PM, Marcelino Borges <[hidden email]> wrote:

Yes, Samson.
I agree with you. According to my intuition, this case can only be handled by considering individuals.

Best regards.

2017-03-22 13:52 GMT-03:00 Samson Tu <[hidden email]>:

On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:

Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.


Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)

Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.

There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.

With best regards,
Samson

_______________________________________________
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: Modeling substances

Igor Toujilov-2

Sometimes we avoid property chains, for example, because they do not coexist with property cardinality restrictions. In this case (if we plan using property cardinality restrictions), we can write it without property chains, and without SWRL:

 

Person and has some (Bicycle and made-of some Steel) SubclassOf has some Steel

 

Cheers,

Igor

 

 
 
Sent: Wednesday, March 22, 2017 at 8:15 PM
From: "Samson Tu" <[hidden email]>
To: "User support for WebProtege and Protege Desktop" <[hidden email]>
Subject: Re: [protege-user] Modeling substances
Maybe I should be more explicit. Using OWL class expressions, your case can be handled without any instance of Steel. The SWRL rule involves individual Person and Bicycle, but not any Steel individual.
 
With best regards,
Samson
 
On Mar 22, 2017, at 12:00 PM, Marcelino Borges <[hidden email]> wrote:
 
Yes, Samson.
I agree with you. According to my intuition, this case can only be handled by considering individuals.
 
Best regards.
 
2017-03-22 13:52 GMT-03:00 Samson Tu <[hidden email]>:
 
On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:
 
Hi.
 
Indeed, you are right, Csongor.
 
However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.
 
For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.
 
The example is a little weird, but it's just to keep our previous example.
 
Best regards.
 
Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)
 
Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.
 
There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.
 
With best regards,
Samson

_______________________________________________
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

_______________________________________________
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: Modeling substances

Marcelino Borges
In reply to this post by samsontu
Hi Sanson.

The singleton pattern seems interesting for my purposes. Is there any reference about this design pattern?
I'm not sure about how to use it in the general case. Would I need to define one single instance for every class in my taxonomy or just for the leafs of the taxonomies?

Using a default instance for representing a more abstract class seems to lead to other problemas. For example, if I need to reclassify it in a more specific class in the future, this would lead to problemas.

Best regards.

2017-03-22 13:52 GMT-03:00 Samson Tu <[hidden email]>:

On Mar 22, 2017, at 5:45 AM, Marcelino Borges <[hidden email]> wrote:

Hi.

Indeed, you are right, Csongor.

However, I think that there are some inferences that we are able to perform only in the level of individuals, so far as I know.

For example, let us consider that someone can infer that: 
if a Person p has a Bicycle b and Bicycle b is made of Steel, them person p has Steel.

The example is a little weird, but it's just to keep our previous example.

Best regards.


Person(?p), Bicycle(?b), has(?p, ?b), (made-of some Steel)(?b) -> (has some Steel)(?p)

Use the Rules tab and Pellet reasoner, not SWRLTab, which does not support class expression as rule predicate.

There may be situations where you need to work with individuals only. One approach is to follow the singleton pattern and use a single individual to stand for a generic instance.

With best regards,
Samson

_______________________________________________
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...