Towards Automatic Extraction of Opinions from Blogs
Abstract: As it has become very easy to publish content online, more
and more people now express their views and personal remarks through blogs. Blogs are highly
subjective and opinionated, but blog content may not
always exclusively consist of opinions. It is common to find blogs containing description of events, factual
information, as well as extracts from news sources and other similar objective
content. Extraction of opinions from blogs is
therefore not trivial, and is an open research problem.
We present Natural Language Processing (NLP) techniques which can be applied to automatic extraction of opinion sentences from blogs. We discuss the challenges in processing blogs given the wide variety of topics and writing styles found in blogs, as well as in the evaluation of results. We also show some potential applications of opinion extraction from blogs.