Location based Twitter Opinion Mining using Common-Sense Information

Abstract

Sentiment analysis research of public information from social networking sites has been increasing immensely in recent years.
Data available at social networking sites is one of the most effective and accurate source to identify the public sentiment of any
product/service. In this paper, we propose a novel localized opinion mining model based on common sense information extracted
from ConceptNet ontology. The proposed methodology allows interpretation and utilization of data extracted from social media site
“Twitter” to identify public opinions. This paper includes location specific, male- female specific and concept specific popularities
of product. All extracted concepts are used to calculate senti_score and to build a machine learning model that classifies the user
opinions as positive or negative.