Encoding sentiments

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Encoding sentiments

Wissik, Tanja

Dear list,

has anybody experiences in encoding sentiments deriving from sentiment analysis or suggestions in how to encode such sentiments?

Thank you in advance!

Best wishes,

Tanja



Dr. Tanja Wissik
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206
http://www.oeaw.ac.at/acdh/

Reply | Threaded
Open this post in threaded view
|

Re: Encoding sentiments

Laurent Romary
Dear Tanja,
On the top of my head I would imagine something like <state type=“sentiment” subtype=“sadness”/>
This could be couple with some standOff annotations on your source text.
Does this make sense?
Beste Grüße,
Laurent


Le 10 sept. 2020 à 11:13, Wissik, Tanja <[hidden email]> a écrit :

Dear list,
has anybody experiences in encoding sentiments deriving from sentiment analysis or suggestions in how to encode such sentiments?
Thank you in advance!
Best wishes,
Tanja


Dr. Tanja Wissik 
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences 
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206 
http://www.oeaw.ac.at/acdh/






Reply | Threaded
Open this post in threaded view
|

Re: Encoding sentiments

Jean-Paul Rehr
In reply to this post by Wissik, Tanja
I suppose the question should be asked: what are the research goals for encoding the  sentiment?

I'm not sure what depth you are looking to go to, nor the physical structure of your text, but I've been encoding "sentiments" of different sorts in inquisitorial records simply using <seg type="some_typology" ana="#some_code"> with @ana referring to a list of codes I've developed specific to my project. Those codes with their descriptions are held in <taxonomy>. The inquisitorial records I work with are quite formulaic, so using <seg>, which suggests structure, to encode the sentiment is a happy coincidence.

On another level, I encode all actors found in these records, and at times I assign an actor the agency of a given emotion using @ana pointing to other codes.

The depositions also have another layer of emotion less easy to detect, which is the sentiment of the context of the deponent giving testimony before inquisitors - at times feelings of fear, relief, rage, avoidance, rushing a story, skipping details, etc. Those I might encode using Laurent Romary's suggestion of <state>. Although I would be inclined to using @ana for the encoding of emotions, not @subtype.

I hope this helps.

Kind regards,
JPR

On Thu, Sep 10, 2020 at 11:13 AM Wissik, Tanja <[hidden email]> wrote:

Dear list,

has anybody experiences in encoding sentiments deriving from sentiment analysis or suggestions in how to encode such sentiments?

Thank you in advance!

Best wishes,

Tanja



Dr. Tanja Wissik
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206
http://www.oeaw.ac.at/acdh/

lou
Reply | Threaded
Open this post in threaded view
|

Re: Encoding sentiments

lou
In reply to this post by Wissik, Tanja
The @ana attribute, globally available, is intended to provide a way of associating any element in your encoded text with some kind of analysis. "Sentiment analysis" as a popular kind of analysis didn't exist when this attribute was invented, but I am pretty sure it is still the right one to use for this purpose. The value of the attribute is a pointer which should indicate some structured representation of the analysis itself -- up to you whether you represent it by a simple list of codes, as in a taxonomy; as a complex bundle of features, using feature structures; or some plain old documentation, using interp or span. These general purpose mechanisms are elaborated upon in chapters 17 and 18 of the Guidelines.

On Thu, 10 Sep 2020 at 10:13, Wissik, Tanja <[hidden email]> wrote:

Dear list,

has anybody experiences in encoding sentiments deriving from sentiment analysis or suggestions in how to encode such sentiments?

Thank you in advance!

Best wishes,

Tanja



Dr. Tanja Wissik
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206
http://www.oeaw.ac.at/acdh/

Reply | Threaded
Open this post in threaded view
|

Re: Encoding sentiments

James Cummings-5
In reply to this post by Wissik, Tanja

Hi Tanja,

I would have encoded these by categorising them in a hierarchical taxonomy of <category> elements and then using the @ana attribute to point to one or more of them. But I like that kind of solution for many similar tasks.

Many thanks,

James 


--

Dr James Cummings, [hidden email]
Senior Lecturer in Late-Medieval Literature and Digital Humanities

School of English, Newcastle University


From: TEI (Text Encoding Initiative) public discussion list <[hidden email]> on behalf of Wissik, Tanja <[hidden email]>
Sent: 10 September 2020 10:13
To: [hidden email] <[hidden email]>
Subject: Encoding sentiments
 

⚠ External sender. Take care when opening links or attachments. Do not provide your login details.

Dear list,

has anybody experiences in encoding sentiments deriving from sentiment analysis or suggestions in how to encode such sentiments?

Thank you in advance!

Best wishes,

Tanja



Dr. Tanja Wissik
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206
http://www.oeaw.ac.at/acdh/

Reply | Threaded
Open this post in threaded view
|

AW: Encoding sentiments

Wissik, Tanja

Dear all,

thank you so much for all your suggestions and input.

Best wishes,
Tanja




Dr. Tanja Wissik
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206
http://www.oeaw.ac.at/acdh/




Von: James Cummings <[hidden email]>
Gesendet: Donnerstag, 10. September 2020 15:05
An: [hidden email]; Wissik, Tanja
Betreff: Re: Encoding sentiments
 

Hi Tanja,

I would have encoded these by categorising them in a hierarchical taxonomy of <category> elements and then using the @ana attribute to point to one or more of them. But I like that kind of solution for many similar tasks.

Many thanks,

James 


--

Dr James Cummings, [hidden email]
Senior Lecturer in Late-Medieval Literature and Digital Humanities

School of English, Newcastle University


From: TEI (Text Encoding Initiative) public discussion list <[hidden email]> on behalf of Wissik, Tanja <[hidden email]>
Sent: 10 September 2020 10:13
To: [hidden email] <[hidden email]>
Subject: Encoding sentiments
 

⚠ External sender. Take care when opening links or attachments. Do not provide your login details.

Dear list,

has anybody experiences in encoding sentiments deriving from sentiment analysis or suggestions in how to encode such sentiments?

Thank you in advance!

Best wishes,

Tanja



Dr. Tanja Wissik
ACDH-CH - Austrian Centre for Digital Humanities and Cultural Heritage
Austrian Academy of Sciences
Sonnenfelsgasse 19 A-1010 Vienna
E-mail: [hidden email]
Tel: + 43 1 51581 - 2206
http://www.oeaw.ac.at/acdh/