DOCUMENT LEVEL CLASSIFICATION USING TWITTER SENTIMENTS

Abstract

Feelings assume an overriding weight in the choices individual make while making purchases.
The expansion in the number of sites has rendered gathering of online audits much easier. We
need a methodology to sift and retrieve relevant information. Information has been gathered
from Text Analysis by Aylien. Information extract using tool Rapidminer to gather and
evaluate tweets. Categorize operator have been utilized to perform document level analysis.
Result concludes that the major shortcoming of document level analysis is unable to known
which feature user is giving opinion.