- Tweet Acts: A Speech Act Classifier for Twitter, Soroush Vosoughi, Deb Ro https://arxiv.org/abs/1605.05156
- "However, there has been very little work done on classifying the pragmatics of tweets"
- how context and intention contribute to meaning
- a set of communicative acts, more commonly known as speech acts.
- noisy and unconventional nature of the language used on the platform
- In this work, we created a supervised speech act classifier for Twitter, using a manually annotated dataset of a few thousand tweets, in order to be better understand the meaning and intention behind tweets and uncover the rich interactions between the users of Twitter.
- Using Searle’s speech act taxonomy (Searle 1976), we established a list of six speech act categories that are commonly seen on Twitter:
- A Assertion (Påstand)
- R Recommendation (Anbefaling)
- E Expression (Uttrykk)
- Q Question (Spørsmål)
- R Request (be om)
- M Miscellaneous (diverse).
- We used Zhao et al.’s (Zhao and Jiang 2011) definitions for topic and type
- A topic is a subject discussed in one or more tweets (e.g., Boston Marathon bombings, Red Sox, etc). The type characterizes the nature of the topic, these are:
- Entity-oriented - Bibliotekets samlinger
- Event-oriented topics - Arrangementer, Tekniske oppdateringer, Undersøkelser, Q & A, (Utstillinger, Åpningstider)
- Long-standing topics - (IK, Open Access) Community, Ansatte, Club.
Dele ALLE tweets i 3
og kombinere med språkhandlinger (gjør digitalt)?
og kombinere med språkhandlinger (gjør digitalt)?
- Påstand - Anbefaling - Uttrykk - Spørsmål - Diverse m/
- Ressurser (samlinger, bøker, artikler, oppslagsverk osv.)
- Eventer (arrangementer/utstillinger/artikler/nettsider/åpningstider)
- Tema (IK og Open Access)
Ingen kommentarer:
Legg inn en kommentar