figshare
Browse
estonianvalence.csv (1.34 MB)

Estonian Valence Corpus / Eesti valentsikorpus

Download (1.34 MB)
dataset
posted on 2023-11-08, 07:59 authored by Hille PajupuuHille Pajupuu, Jaan Pajupuu, Rene Altrov, Kairi Tamuri

The Estonian Valence Corpus consists of paragraphs (4,088) from newspaper articles and comments in "Postimees" and "Õhtuleht," whose emotional tone (positive, negative, mixed, neutral) has been determined by readers. The dominant opinion method was employed to determine the tone (Pennebaker, James W.; Mayne, Tracy J.; Francis, Martha E. 1997. Linguistic predictors of adaptive bereavement. Journal of Personality and Social Psychology 72(4), 863-871).

The Estonian Valence Corpus is primarily intended for training statistical models but can also be used for other purposes.

"Postimees" sections/categories: Opinion (ARVAMUS), Culture (KULTUUR), Sports (SPORT), International (VÄLISMAA), Crime (KRIMI), Estonia (EESTI), comments on Estonia (KOMM-P-EESTI)

"Õhtuleht" sections/categories: Life (ELU-O), comments on Life (KOMM-O-ELU)

For more information about the corpus: Pajupuu, Hille; Altrov, Rene; Pajupuu, Jaan (2016). Identifying polarity in different text types. Folklore: Electronic Journal of Folklore, 64, 25−42. https://doi.org/10.7592/FEJF2016.64.polarity.

***

Eesti valentsikorpus koosneb "Postimehe" ja "Õhtulehe" artiklite ja kommentaaride ortograafilistest lõikudest (4088), mille emotsionaalsuse (positiivne, negatiivne, vastuoluline, neutraalne) on määranud lugejad. Kasutatud on domineeriva arvamuse meetodit (Pennebaker, James W.; Mayne, Tracy J.; Francis, Martha E. 1997. Linguistic predictors of adaptive bereavement. Journal of Personality and Social Psychology 72(4), 863-871).

Valentsikorpus on mõeldud eeskätt statistiliste mudelite treenimiseks, kuid seda saab kasutada ka muudel eesmärkidel.

"Postimehe" rubriigid: ARVAMUS, KULTUUR, SPORT, VÄLISMAA, KRIMI, EESTI, KOMM-P-EESTI (Eesti kommentaarid)

"Õhtulehe" rubriigid: ELU-O (Elu), KOMM-O-ELU (Elu kommentaarid)

Korpusest täpsemalt: Pajupuu, Hille; Altrov, Rene; Pajupuu, Jaan (2016). Identifying polarity in different text types. Folklore: Electronic Journal of Folklore, 64, 25−42. https://doi.org/10.7592/FEJF2016.64.polarity.


Funding

EKT1: Kõne ja teksti emotsionaalsuse statistilised mudelid (2011-2014)

History

Usage metrics

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC