MKLab-ITI / greek-sentiment-lexicon

A lexicon to be used for sentiment analysis in Greek.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

greek-sentiment-lexicon

This work is superseded by the following paper: Tsakalidis, A., Papadopoulos, S., Voskaki, R., Ioannidou, K., Boididou, C., Cristea, A. I., Liakata, M. & Kompatsiaris, Y. (2018). Building and evaluating resources for sentiment analysis in the Greek language. Language resources and evaluation, 52(4), 1021-1044. Please read the paper and download the relevant resources from the following link.

A lexicon to be used for sentiment analysis in Greek.

This lexicon was created with the support of the EC-funded FP7 Project SocialSensor. Further information on the creation of the lexicon is available on Deliverable D2.3: Social Stream Mining Framework (p.66) of the project and on the respective project page.

Each row of the lexicon is tab-separated and corresponds to the following fields:

  • Term: Greek term to which the following annotations refer. In the cases of adjectives, all three genders (male, female, neutral) are implied with the provision of the suffixes (-ος -η -ο). In some cases, the terms refer to components of larger words. Those terms end with hyphen (-).
  • POS_x: Part of Speech tag assigned by annotator x (x=1,2,3,4). In a few complicated cases, there is no unanimous annotator agreement on the POS of a term.
  • Subjectivity_x: Subjectivity class assigned by annotator x. Possible values are Objective (OBJ), Weakly Subjective (SUBJ-) and Strongly Subjective (SUBJ+).
  • Polarity_x: Polarity class assigned by annotator x. Possible values are Positive (POS), Negative (NEG), Both positive and negative (BOTH) and Neutral (N/A).
  • Anger_x: Anger score assigned by annotator x. Possible values are between 1 and 5, and N/A.
  • Disgust_x: Disgust score assigned by annotator x. Possible values are between 1 and 5, and N/A.
  • Fear_x: Fear score assigned by annotator x. Possible values are between 1 and 5, and N/A.
  • Happiness_x: Happiness score assigned by annotator x. Possible values are between 1 and 5, and N/A.
  • Sadness_x: Sadness score assigned by annotator x. Possible values are between 1 and 5, and N/A.
  • Surprise_x: Surprise score assigned by annotator x. Possible values are between 1 and 5, and N/A.
  • Additional_x: Other terms that annotator x considered as relevant with this term.
  • Comments_x: Free-text comments to clarify the usage of the term.

N/A (Not Applicable) value is used in cases that the annotator considers that no value is appropriate for the term or in cases where he/she was not confident on the appropriate value.

Acknoweldgements: The lexicon was generated by Adam Tsakalidis (CERTH-ITI, now University of Warwick) in collaboration with Symeon Papadopoulos (CERTH-ITI) and with the contribution of Ourania Voskaki (Centre for Greek Language) and Kyriaki Ioannidou (Centre for Greek Language) and Christina Boididou (CERTH-ITI). For feedback and suggestions, please get in touch with adam.tsakalidis@gmail.com and papadop@iti.gr.

About

A lexicon to be used for sentiment analysis in Greek.

License:Apache License 2.0