Pregled bibliografske jedinice broj: 638878
Sentiscope: A System for Sentiment Analysis in Daily Horoscopes
Sentiscope: A System for Sentiment Analysis in Daily Horoscopes // Formalising Natural Languages with NooJ / Donabédian, Anaïd ; Khurshudian, Victoria ; Silberztein, Max (ur.).
Newcastle upon Tyne: Cambridge Scholars Publishing, 2013. str. 173-181
CROSBI ID: 638878 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Sentiscope: A System for Sentiment Analysis in Daily Horoscopes
Autori
Merkler, Danijela ; Agić, Željko
Vrsta, podvrsta i kategorija rada
Poglavlja u knjigama, znanstveni
Knjiga
Formalising Natural Languages with NooJ
Urednik/ci
Donabédian, Anaïd ; Khurshudian, Victoria ; Silberztein, Max
Izdavač
Cambridge Scholars Publishing
Grad
Newcastle upon Tyne
Godina
2013
Raspon stranica
173-181
ISBN
978-1-4438-4733-9
Ključne riječi
sentiment analysis, narrow domain, rule-based system
Sažetak
We present Sentiscope – a prototype system for collecting, sentiment annotation and visualization of daily horoscopes from news portals written in Croatian. System architecture and evaluation is provided and the system itself is made available to users on the web. The sentiment analysis component is based on rules implemented as finite state transducer cascades in NooJ development environment.
Izvorni jezik
Engleski
Znanstvena područja
Informacijske i komunikacijske znanosti
POVEZANOST RADA
Projekti:
130-1300646-1776 - Računalna sintaksa hrvatskoga jezika (Dovedan Han, Zdravko, MZOS ) ( CroRIS)
Ustanove:
Filozofski fakultet, Zagreb
Profili:
Željko Agić
(autor)
