Abstract
An information extraction system that extracts 'sentiment' words in unrestricted Arabic texts is described. Earlier work on the automatic sentiment analysis in English and Chinese texts has been adapted and extended to the morphologically richer Arabic texts. A method for automatically and unambiguously extracting sentiment-bearing patterns from texts in Arabic is described. A list of sentiment words is provided and sentiment bearing excerpts are extracted from time-stamped news wire: The frequency of sentiment words per unit time is then plotted for 'visualising' changes in sentiments.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have