You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jena.apache.org by javed khan <ja...@gmail.com> on 2018/01/21 13:50:49 UTC
Rules and machine learning
Hello
I am not sure if the question is related to the jena group but I will
appreciate the answer.
I want to ask is it possible we take the functionality of machine learning
techniques (bayes algorithm, decision tree etc) using semantic web rules. I
dont know much about machine learning but I know it makes prediction based
on past experience/past data.
Like we provide set of rules based on past data (if this, then that) and
make predictions/optimizations. For instance, we want to make bug
predictions in a software using Semantic rules, so is it possible??
Thank you
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
Re: Rules and machine learning
Posted by javed khan <ja...@gmail.com>.
Thank you Lorenz.
And sorry for the off-topic query.
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
On Tue, Jan 23, 2018 at 10:59 AM, Lorenz Buehmann <
buehmann@informatik.uni-leipzig.de> wrote:
> > The
> > following is a paper which used "inconsistency detection" but not used
> ML.
> I know the authors and the paper. Your statement is not correct, in fact
> they used machine learning to find new schema axioms to be added the
> DBpedia ontology, in particular disjointness axioms which are quite
> often a "good" source for inconsistency.
>
> There are also other groups that already used machine learning to
> "replace" standard reasoning procedures.
>
> But again, it's too much off-topic here.
>
>
> On 22.01.2018 17:56, javed khan wrote:
> > Martin, thanks a lot. Its very useful for me..
> >
> > I think it will be possible also to predict inconsistencies in ontologies
> > (via machine learning). It could be my research project which is about to
> > start but the problem is I cant find anything related on the web. The
> > following is a paper which used "inconsistency detection" but not used
> ML.
> >
> > [1]
> > https://hpi.de/fileadmin/user_upload/fachgebiete/meinel/
> papers/Web_3.0/2012_Toepper_ISEM.pdf
> >
> >
> >
> > <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> > Virus-free.
> > www.avast.com
> > <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> > <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> >
> > On Mon, Jan 22, 2018 at 6:52 PM, Martin Vachovski <
> > Martin.Vachovski@skytek.com> wrote:
> >
> >> Hi all,
> >>
> >> I have seen some papers on "ontology matching"
> >> which is to say- apply a ML algorithm in order to "map"
> >> the semantics of two different ontologies which apply to the same object
> >>
> >> https://homes.cs.washington.edu/~pedrod/papers/hois.pdf
> >> http://disi.unitn.it/~p2p/RelatedWork/Matching/0411csit10.pdf
> >>
> >> While the examples are not exactly the ones seek by the question, they
> >> show that the idea of combining of ML and semantic data storage is not
> new
> >> Hope that points towards the right direction
> >>
> >> Cheers
> >> Martin
> >>
> >>
> >> ________________________________________
> >> From: javed khan <ja...@gmail.com>
> >> Sent: Monday, January 22, 2018 3:12 PM
> >> To: users@jena.apache.org
> >> Subject: Re: Rules and machine learning
> >>
> >> Thank you Lorenz.. Yes rules can not be consider machine learning as
> its a
> >> kind of hard coding and machine will not learn by itself..
> >>
> >>
> >>
> >> <https://www.avast.com/sig-email?utm_medium=email&utm_
> >> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> >> Virus-free.
> >> www.avast.com
> >> <https://www.avast.com/sig-email?utm_medium=email&utm_
> >> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> >> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> >>
> >> On Mon, Jan 22, 2018 at 7:35 AM, Lorenz Buehmann <
> >> buehmann@informatik.uni-leipzig.de> wrote:
> >>
> >>> I've just some very minimal experience in machine learning and rule
> >>> processing...
> >>>
> >>> The important keywords here are "machine" and "learning" - if you
> >>> provide a set of rules, then there was no learning. Except by the human
> >>> who used his/her knowledge to make the rules - if it's done by a
> >>> "machine", then you can call this machine learning of such rules (e.g.
> >>> rule induction) and use the rules to "infer" data - not predict. The
> >>> rules are just a (human-readable) way to encode the machine learning
> >> model.
> >>> But, it's off-topic for sure, thus, I will not go further into details.
> >>>
> >>>
> >>> Lorenz
> >>>
> >>>
> >>> On 21.01.2018 14:50, javed khan wrote:
> >>>> Hello
> >>>>
> >>>> I am not sure if the question is related to the jena group but I will
> >>>> appreciate the answer.
> >>>>
> >>>> I want to ask is it possible we take the functionality of machine
> >>> learning
> >>>> techniques (bayes algorithm, decision tree etc) using semantic web
> >>> rules. I
> >>>> dont know much about machine learning but I know it makes prediction
> >>> based
> >>>> on past experience/past data.
> >>>>
> >>>> Like we provide set of rules based on past data (if this, then that)
> >> and
> >>>> make predictions/optimizations. For instance, we want to make bug
> >>>> predictions in a software using Semantic rules, so is it possible??
> >>>>
> >>>> Thank you
> >>>>
> >>>> <https://www.avast.com/sig-email?utm_medium=email&utm_
> >>> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> >>>> Virus-free.
> >>>> www.avast.com
> >>>> <https://www.avast.com/sig-email?utm_medium=email&utm_
> >>> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> >>>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> >>>>
> >>>
> >>>
>
>
Re: Rules and machine learning
Posted by Lorenz Buehmann <bu...@informatik.uni-leipzig.de>.
> The
> following is a paper which used "inconsistency detection" but not used ML.
I know the authors and the paper. Your statement is not correct, in fact
they used machine learning to find new schema axioms to be added the
DBpedia ontology, in particular disjointness axioms which are quite
often a "good" source for inconsistency.
There are also other groups that already used machine learning to
"replace" standard reasoning procedures.
But again, it's too much off-topic here.
On 22.01.2018 17:56, javed khan wrote:
> Martin, thanks a lot. Its very useful for me..
>
> I think it will be possible also to predict inconsistencies in ontologies
> (via machine learning). It could be my research project which is about to
> start but the problem is I cant find anything related on the web. The
> following is a paper which used "inconsistency detection" but not used ML.
>
> [1]
> https://hpi.de/fileadmin/user_upload/fachgebiete/meinel/papers/Web_3.0/2012_Toepper_ISEM.pdf
>
>
>
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> Virus-free.
> www.avast.com
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>
> On Mon, Jan 22, 2018 at 6:52 PM, Martin Vachovski <
> Martin.Vachovski@skytek.com> wrote:
>
>> Hi all,
>>
>> I have seen some papers on "ontology matching"
>> which is to say- apply a ML algorithm in order to "map"
>> the semantics of two different ontologies which apply to the same object
>>
>> https://homes.cs.washington.edu/~pedrod/papers/hois.pdf
>> http://disi.unitn.it/~p2p/RelatedWork/Matching/0411csit10.pdf
>>
>> While the examples are not exactly the ones seek by the question, they
>> show that the idea of combining of ML and semantic data storage is not new
>> Hope that points towards the right direction
>>
>> Cheers
>> Martin
>>
>>
>> ________________________________________
>> From: javed khan <ja...@gmail.com>
>> Sent: Monday, January 22, 2018 3:12 PM
>> To: users@jena.apache.org
>> Subject: Re: Rules and machine learning
>>
>> Thank you Lorenz.. Yes rules can not be consider machine learning as its a
>> kind of hard coding and machine will not learn by itself..
>>
>>
>>
>> <https://www.avast.com/sig-email?utm_medium=email&utm_
>> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
>> Virus-free.
>> www.avast.com
>> <https://www.avast.com/sig-email?utm_medium=email&utm_
>> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>>
>> On Mon, Jan 22, 2018 at 7:35 AM, Lorenz Buehmann <
>> buehmann@informatik.uni-leipzig.de> wrote:
>>
>>> I've just some very minimal experience in machine learning and rule
>>> processing...
>>>
>>> The important keywords here are "machine" and "learning" - if you
>>> provide a set of rules, then there was no learning. Except by the human
>>> who used his/her knowledge to make the rules - if it's done by a
>>> "machine", then you can call this machine learning of such rules (e.g.
>>> rule induction) and use the rules to "infer" data - not predict. The
>>> rules are just a (human-readable) way to encode the machine learning
>> model.
>>> But, it's off-topic for sure, thus, I will not go further into details.
>>>
>>>
>>> Lorenz
>>>
>>>
>>> On 21.01.2018 14:50, javed khan wrote:
>>>> Hello
>>>>
>>>> I am not sure if the question is related to the jena group but I will
>>>> appreciate the answer.
>>>>
>>>> I want to ask is it possible we take the functionality of machine
>>> learning
>>>> techniques (bayes algorithm, decision tree etc) using semantic web
>>> rules. I
>>>> dont know much about machine learning but I know it makes prediction
>>> based
>>>> on past experience/past data.
>>>>
>>>> Like we provide set of rules based on past data (if this, then that)
>> and
>>>> make predictions/optimizations. For instance, we want to make bug
>>>> predictions in a software using Semantic rules, so is it possible??
>>>>
>>>> Thank you
>>>>
>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_
>>> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
>>>> Virus-free.
>>>> www.avast.com
>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_
>>> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
>>>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>>>>
>>>
>>>
Re: Rules and machine learning
Posted by javed khan <ja...@gmail.com>.
Martin, thanks a lot. Its very useful for me..
I think it will be possible also to predict inconsistencies in ontologies
(via machine learning). It could be my research project which is about to
start but the problem is I cant find anything related on the web. The
following is a paper which used "inconsistency detection" but not used ML.
[1]
https://hpi.de/fileadmin/user_upload/fachgebiete/meinel/papers/Web_3.0/2012_Toepper_ISEM.pdf
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
On Mon, Jan 22, 2018 at 6:52 PM, Martin Vachovski <
Martin.Vachovski@skytek.com> wrote:
> Hi all,
>
> I have seen some papers on "ontology matching"
> which is to say- apply a ML algorithm in order to "map"
> the semantics of two different ontologies which apply to the same object
>
> https://homes.cs.washington.edu/~pedrod/papers/hois.pdf
> http://disi.unitn.it/~p2p/RelatedWork/Matching/0411csit10.pdf
>
> While the examples are not exactly the ones seek by the question, they
> show that the idea of combining of ML and semantic data storage is not new
> Hope that points towards the right direction
>
> Cheers
> Martin
>
>
> ________________________________________
> From: javed khan <ja...@gmail.com>
> Sent: Monday, January 22, 2018 3:12 PM
> To: users@jena.apache.org
> Subject: Re: Rules and machine learning
>
> Thank you Lorenz.. Yes rules can not be consider machine learning as its a
> kind of hard coding and machine will not learn by itself..
>
>
>
> <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> Virus-free.
> www.avast.com
> <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>
> On Mon, Jan 22, 2018 at 7:35 AM, Lorenz Buehmann <
> buehmann@informatik.uni-leipzig.de> wrote:
>
> > I've just some very minimal experience in machine learning and rule
> > processing...
> >
> > The important keywords here are "machine" and "learning" - if you
> > provide a set of rules, then there was no learning. Except by the human
> > who used his/her knowledge to make the rules - if it's done by a
> > "machine", then you can call this machine learning of such rules (e.g.
> > rule induction) and use the rules to "infer" data - not predict. The
> > rules are just a (human-readable) way to encode the machine learning
> model.
> >
> > But, it's off-topic for sure, thus, I will not go further into details.
> >
> >
> > Lorenz
> >
> >
> > On 21.01.2018 14:50, javed khan wrote:
> > > Hello
> > >
> > > I am not sure if the question is related to the jena group but I will
> > > appreciate the answer.
> > >
> > > I want to ask is it possible we take the functionality of machine
> > learning
> > > techniques (bayes algorithm, decision tree etc) using semantic web
> > rules. I
> > > dont know much about machine learning but I know it makes prediction
> > based
> > > on past experience/past data.
> > >
> > > Like we provide set of rules based on past data (if this, then that)
> and
> > > make predictions/optimizations. For instance, we want to make bug
> > > predictions in a software using Semantic rules, so is it possible??
> > >
> > > Thank you
> > >
> > > <https://www.avast.com/sig-email?utm_medium=email&utm_
> > source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> > > Virus-free.
> > > www.avast.com
> > > <https://www.avast.com/sig-email?utm_medium=email&utm_
> > source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> > > <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> > >
> >
> >
> >
>
Re: Rules and machine learning
Posted by Martin Vachovski <Ma...@skytek.com>.
Hi all,
I have seen some papers on "ontology matching"
which is to say- apply a ML algorithm in order to "map"
the semantics of two different ontologies which apply to the same object
https://homes.cs.washington.edu/~pedrod/papers/hois.pdf
http://disi.unitn.it/~p2p/RelatedWork/Matching/0411csit10.pdf
While the examples are not exactly the ones seek by the question, they show that the idea of combining of ML and semantic data storage is not new
Hope that points towards the right direction
Cheers
Martin
________________________________________
From: javed khan <ja...@gmail.com>
Sent: Monday, January 22, 2018 3:12 PM
To: users@jena.apache.org
Subject: Re: Rules and machine learning
Thank you Lorenz.. Yes rules can not be consider machine learning as its a
kind of hard coding and machine will not learn by itself..
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
On Mon, Jan 22, 2018 at 7:35 AM, Lorenz Buehmann <
buehmann@informatik.uni-leipzig.de> wrote:
> I've just some very minimal experience in machine learning and rule
> processing...
>
> The important keywords here are "machine" and "learning" - if you
> provide a set of rules, then there was no learning. Except by the human
> who used his/her knowledge to make the rules - if it's done by a
> "machine", then you can call this machine learning of such rules (e.g.
> rule induction) and use the rules to "infer" data - not predict. The
> rules are just a (human-readable) way to encode the machine learning model.
>
> But, it's off-topic for sure, thus, I will not go further into details.
>
>
> Lorenz
>
>
> On 21.01.2018 14:50, javed khan wrote:
> > Hello
> >
> > I am not sure if the question is related to the jena group but I will
> > appreciate the answer.
> >
> > I want to ask is it possible we take the functionality of machine
> learning
> > techniques (bayes algorithm, decision tree etc) using semantic web
> rules. I
> > dont know much about machine learning but I know it makes prediction
> based
> > on past experience/past data.
> >
> > Like we provide set of rules based on past data (if this, then that) and
> > make predictions/optimizations. For instance, we want to make bug
> > predictions in a software using Semantic rules, so is it possible??
> >
> > Thank you
> >
> > <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> > Virus-free.
> > www.avast.com
> > <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> > <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> >
>
>
>
Re: Rules and machine learning
Posted by javed khan <ja...@gmail.com>.
Thank you Lorenz.. Yes rules can not be consider machine learning as its a
kind of hard coding and machine will not learn by itself..
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
On Mon, Jan 22, 2018 at 7:35 AM, Lorenz Buehmann <
buehmann@informatik.uni-leipzig.de> wrote:
> I've just some very minimal experience in machine learning and rule
> processing...
>
> The important keywords here are "machine" and "learning" - if you
> provide a set of rules, then there was no learning. Except by the human
> who used his/her knowledge to make the rules - if it's done by a
> "machine", then you can call this machine learning of such rules (e.g.
> rule induction) and use the rules to "infer" data - not predict. The
> rules are just a (human-readable) way to encode the machine learning model.
>
> But, it's off-topic for sure, thus, I will not go further into details.
>
>
> Lorenz
>
>
> On 21.01.2018 14:50, javed khan wrote:
> > Hello
> >
> > I am not sure if the question is related to the jena group but I will
> > appreciate the answer.
> >
> > I want to ask is it possible we take the functionality of machine
> learning
> > techniques (bayes algorithm, decision tree etc) using semantic web
> rules. I
> > dont know much about machine learning but I know it makes prediction
> based
> > on past experience/past data.
> >
> > Like we provide set of rules based on past data (if this, then that) and
> > make predictions/optimizations. For instance, we want to make bug
> > predictions in a software using Semantic rules, so is it possible??
> >
> > Thank you
> >
> > <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> > Virus-free.
> > www.avast.com
> > <https://www.avast.com/sig-email?utm_medium=email&utm_
> source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> > <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> >
>
>
>
Re: Rules and machine learning
Posted by Lorenz Buehmann <bu...@informatik.uni-leipzig.de>.
I've just some very minimal experience in machine learning and rule
processing...
The important keywords here are "machine" and "learning" - if you
provide a set of rules, then there was no learning. Except by the human
who used his/her knowledge to make the rules - if it's done by a
"machine", then you can call this machine learning of such rules (e.g.
rule induction) and use the rules to "infer" data - not predict. The
rules are just a (human-readable) way to encode the machine learning model.
But, it's off-topic for sure, thus, I will not go further into details.
Lorenz
On 21.01.2018 14:50, javed khan wrote:
> Hello
>
> I am not sure if the question is related to the jena group but I will
> appreciate the answer.
>
> I want to ask is it possible we take the functionality of machine learning
> techniques (bayes algorithm, decision tree etc) using semantic web rules. I
> dont know much about machine learning but I know it makes prediction based
> on past experience/past data.
>
> Like we provide set of rules based on past data (if this, then that) and
> make predictions/optimizations. For instance, we want to make bug
> predictions in a software using Semantic rules, so is it possible??
>
> Thank you
>
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
> Virus-free.
> www.avast.com
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>