Emotion annotation scenario #43
Comments
|
Hi @luigi-asprino , I would like to add a tag referring the whole document (specifically an overall document sentiment value). Should I create a range w.r.t. the whole document or do you suggest another approach? |
|
Hi @alessioBosca, What kind of annotation would you like to add? Could you give some examples? Thanks! Luigi |
|
Hi @luigi-asprino, I would like to add an overall document sentiment value, I guess the value could be expressed like for the span "semiotics:denotes" : { however I'm not sure how to replace "@type" : "earmark:PointerRange", and begin/end . I can refer to the begin / end of the document but maybe there is a better way to express it |
|
you can use semiotics:denotes as predicate, but I would rather use the docuverse (ex:docuverse in the example) as the subject of the triple to refer to the whole document instead of the pointer ranges. |
|
Hello, the docuverse must be the subject indeed, but the semiotics:denotes is typically used to denote what the document refers to, e.g. “I love you” denotes an event in which e.g. a human is the agent/subject of loving another one (the patient/object of loving).
Positivity (marl:Positive) is only indirectly denoted by the text, since it is averaged across the multiple sentiment scores computed by the algorithm (or otherwise, but still computed), so I’d use a more specific predicate, e.g. the one used in Sentilo: <http://www.ontologydesignpatterns.org/ont/sentilo.owl#hasAvgScore>.
Aldo
… On 19 Jan 2022, at 10:55, luigi-asprino ***@***.***> wrote:
you can use semiotics:denotes as predicate, but I would rather use the docuverse (ex:docuverse in the example) as the subject of the triple to refer to the whole document instead of the pointer ranges.
I'm going to update the example above to reflect this annotation.
—
Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONFMTPVH4AVPHBMO3G3UW2C73ANCNFSM4YBBEFVQ>.
You are receiving this because you are subscribed to this thread.
|
|
Ok great. I will update the example. Any chance you have the source code of the sentilo ontology?
Luigi
… On 19 Jan 2022, at 12:41, aldogangemi ***@***.***> wrote:
Hello, the docuverse must be the subject indeed, but the semiotics:denotes is typically used to denote what the document refers to, e.g. “I love you” denotes an event in which e.g. a human is the agent/subject of loving another one (the patient/object of loving).
Positivity (marl:Positive) is only indirectly denoted by the text, since it is averaged across the multiple sentiment scores computed by the algorithm (or otherwise, but still computed), so I’d use a more specific predicate, e.g. the one used in Sentilo: <http://www.ontologydesignpatterns.org/ont/sentilo.owl#hasAvgScore>.
Aldo
> On 19 Jan 2022, at 10:55, luigi-asprino ***@***.***> wrote:
>
>
> you can use semiotics:denotes as predicate, but I would rather use the docuverse (ex:docuverse in the example) as the subject of the triple to refer to the whole document instead of the pointer ranges.
> I'm going to update the example above to reflect this annotation.
>
> —
> Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONFMTPVH4AVPHBMO3G3UW2C73ANCNFSM4YBBEFVQ>.
> You are receiving this because you are subscribed to this thread.
>
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications on the go with GitHub Mobile for iOS or Android.
You are receiving this because you were mentioned.
|
|
Sure, that’s the sentilo.owl link:
http://ontologydesignpatterns.org/ont/sentilo.owl <http://ontologydesignpatterns.org/ont/sentilo.owl>
Aldo
… On 19 Jan 2022, at 12:44, luigi-asprino ***@***.***> wrote:
Ok great. I will update the example. Any chance you have the source code of the sentilo ontology?
Luigi
> On 19 Jan 2022, at 12:41, aldogangemi ***@***.***> wrote:
>
>
> Hello, the docuverse must be the subject indeed, but the semiotics:denotes is typically used to denote what the document refers to, e.g. “I love you” denotes an event in which e.g. a human is the agent/subject of loving another one (the patient/object of loving).
> Positivity (marl:Positive) is only indirectly denoted by the text, since it is averaged across the multiple sentiment scores computed by the algorithm (or otherwise, but still computed), so I’d use a more specific predicate, e.g. the one used in Sentilo: <http://www.ontologydesignpatterns.org/ont/sentilo.owl#hasAvgScore>.
> Aldo
>
>
> > On 19 Jan 2022, at 10:55, luigi-asprino ***@***.***> wrote:
> >
> >
> > you can use semiotics:denotes as predicate, but I would rather use the docuverse (ex:docuverse in the example) as the subject of the triple to refer to the whole document instead of the pointer ranges.
> > I'm going to update the example above to reflect this annotation.
> >
> > —
> > Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONFMTPVH4AVPHBMO3G3UW2C73ANCNFSM4YBBEFVQ>.
> > You are receiving this because you are subscribed to this thread.
> >
>
> —
> Reply to this email directly, view it on GitHub, or unsubscribe.
> Triage notifications on the go with GitHub Mobile for iOS or Android.
> You are receiving this because you were mentioned.
>
—
Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONC5HVUF6MC5HP4XZWDUW2P2BANCNFSM4YBBEFVQ>.
You are receiving this because you commented.
|
|
Great! Thank you |
|
@luigi-asprino @aldogangemi thanks for the update and examples Alessio |
|
Ah giusto una issue tecnica: se il valore può essere un individuo come marl:Positive, allora qui c’è bisogno di una ObjectProperty, mentre sentilo:hasAVGScore è una DatatypeProperty (il range è xsd:double). Forse si potrebbe aggiungere una ObjectProperty a sentilo.owl, ex. sentilo:scoreType.
… On 19 Jan 2022, at 17:22, Alessio Bosca ***@***.***> wrote:
@luigi-asprino <https://github.com/luigi-asprino> @aldogangemi <https://github.com/aldogangemi> thanks for the update and examples
Alessio
—
Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONFV5YVTM5N6CQ7FHW3UW3QMBANCNFSM4YBBEFVQ>.
You are receiving this because you were mentioned.
|
|
Ok, shall we use sentilo:scoreType? |
|
OK, I’ll do right away. And sorry for the Italian switch ...
… On 19 Jan 2022, at 18:48, luigi-asprino ***@***.***> wrote:
Ok, shall we use sentilo:scoreType?
Could you add such object property to the sentilo ontology (I don't have access to the ODP's FTP server anymore).
—
Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONGGVOGU5RCRE2AKNELUW32O7ANCNFSM4YBBEFVQ>.
You are receiving this because you were mentioned.
|
|
Hi @luigi-asprino and @aldogangemi, I would like to combine scoreType with avgScore (in order to provede a confidence value), I was thinking something like "sentilo:scoreType" : [{ would that be OK? |
|
Sure Alessio, that’s good.
Aldo
… On 24 Jan 2022, at 15:38, Alessio Bosca ***@***.***> wrote:
Hi @luigi-asprino <https://github.com/luigi-asprino> and @aldogangemi <https://github.com/aldogangemi>, I would like to combine scoreType with avgScore (in order to provede a confidence value), I was thinking something like
"sentilo:scoreType" : [{
***@***.*** <https://github.com/id>": "marl:Positive",
"sentilo:avgScore": 0.6
},
{
***@***.*** <https://github.com/type>" : "emotion:Calmness",
"sentilo:avgScore": 0.82
}]
would that be OK?
—
Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB52ONB57IN4SMGH2EPVC3LUXVP5JANCNFSM4YBBEFVQ>.
You are receiving this because you were mentioned.
|
|
Thanks Aldo |
|
Hi, I have updated the example. |
Depends on
#30
Description
Emotion detector annotates a piece of text with emotions of the Pluchick ontology.
Source
WP3-WP6 working group
Competency Questions
cf. #30
Ontology/Ontology Design Patterns Relevant for the scenario
cf. #30
MARL Opinion Ontology for representing polarity
SPARQL Questions
cf. #30
Toy Dataset/SPARQL Endpoint
The text was updated successfully, but these errors were encountered: