[Corpora-List] EmoText - Software for opinion mining and lexical affect sensing

[Justin Washtell]

Alexander, Yorick,



I think we all agree that for correctly addressing sentiment etc, pragmatic (i.e. such as might include demographic) information plays an important role. Not just for correctly handling lexis ("rubbish", "trash" etc) but, more subtly, stuff like the kind of metaphors people use ("it's the Chicken Tikka Masala of mobile phones") and their prior attitudes ("it was exactly what you'd expect from a Hollywood movie"). And I don't think anybody would disagree that garnering more data, and extralinguistic features, can help to improve our chances in the long run. These are very important observations which Alexander is astute and right to make. However - and at the risk of tempting fate - I also think most of us probably agree that if the present state-of-the-art cannot correctly deal with relatively straightforward compositional cases of common language, then no amount of tuning on the data is going to solve the problem: rather it suggests that something about our base model is inadequate.



Justin Washtell

University of Leeds



________________________________
From: corpora-bounces at uib.no [corpora-bounces at uib.no] On Behalf Of Yorick Wilks [Y.Wilks at dcs.shef.ac.uk]
Sent: 22 December 2011 18:26
To: Alexander Osherenko
Cc: corpora at uib.no
Subject: Re: [Corpora-List] EmoText - Software for opinion mining and lexical affect sensing

Im not sure I understand you but I think you agree with me. There is nothing remotely geographical about the characterization of the language of truck drivers, physicists or miners!
YW

On 22 Dec 2011, at 18:15, Alexander Osherenko wrote:

The initial question was: is it necessary to compose separate datasets for opinion mining in English British, in American English, in Australian English etc or there is some sort of universal dataset that can be used for opinion mining in every English. As far as I understood, Justin thinks there is the universal dataset. I am convinced that only language consideration is not enough and there must be other factors that influence feature extraction. I assumed this influence can be represented by geographical reasons. Moreover, I supposed that weights of features in classifiers are more appropriate to explain pragmatic content of a text and not semantic.

2011/12/22 Yorick Wilks <Y.Wilks at dcs.shef.ac.uk<mailto:Y.Wilks at dcs.shef.ac.uk>>
Surely, the point this interchange has reached has nothing to do with country or variants of a language of the kind a linguist is interested in; for "American truck driver" substitute "local physicist". This has gone into areas where linguistics and NLP have nothing to say if "comprehension" is extended to mean "understands mathematics".
Yorick Wilks


On 22 Dec 2011, at 16:26, Alexander Osherenko wrote:

Justin,

I mean more the distinction between pragmatics and semantics (http://oxfordbibliographiesonline.com/view/document/obo-9780195396577/obo-9780195396577-0146.xml#obo-9780195396577-0146-div1-0010)

AO

2011/12/21 Justin Washtell <lec3jrw at leeds.ac.uk<mailto:lec3jrw at leeds.ac.uk>>
Hello Alexander,

>> As a human, and an Englishman, I expect I can understand and fairly judge the sentiment of most reviews written by, say, an American truck driver, without undue reprogramming. Is this really an unrealistic goal for our algorithms? And I wonder, is mastering a highly restricted style or register a necessary step in that direction... or is it in fact a detour.
> As a human and as an Englishman, you learned to recognize particular words of English language. Now you understand English in every country but you can't comprehend it. Understanding is only the first step of cognition, comprehension takes much more time and energy. Or can you explain the most severe problems of American truck drivers nowadays? Or tell me what problems you would discuss with an American truck driver? In terms of data mining, it means: you know what features you have in a dataset but you don't know their weights. In my opinion, if you want to learn "weights" you have to live in the country and tune the weights.

I was referring only to detecting coarse-grained sentiment, not to appreciating the detailed tribulations of the author's lifestyle, but you make a point :-) The distinction you draw between understanding and comprehension is an unfamiliar one to me. I'm not sure whether it is more akin to the distinction I would make between "recognition" and "understanding", or between "undestanding" and "experiencing" (though both of which I concede are fuzzy at best and lure us beyond the pretentiously tidy realms of NLP and into the free-for-all that is philosophy). In any case, please re-interpret my use of the word to mean something closer to what you would call "comprehension".

Justin

_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no<mailto:Corpora at uib.no>
http://mailman.uib.no/listinfo/corpora


_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no<mailto:Corpora at uib.no>
http://mailman.uib.no/listinfo/corpora


_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora