Enhanced automatic extraction of target based multi word expression with clustering technique for opinion mining

Abstract

Information Extraction IE is an automatic extraction of required text from semistructured or unstructured environments This text data is categorized into objective and subjective sentences An objective sentence contains factual that is true information about the world and subjective sentence contains some personal opinions or views Retrieving information from subjective sentence is an emerging research trend Opinion Mining is the study of users opinion about a product and/or service from plain text which is either done at the sentence document or target level This research concentrates on target level also known as aspect or feature Opinion Mining from different customer reviews Here target level extracts the detailed information about the targets or features of the product and services It deals with the opinions of entities andor targets and their corresponding polarities such as positive negative or neutral Usually two types of targets are present in the customer reviews namely Explicit directly mentioning the targets and Implicit indirectly mentioning the targets Finetuned opinion extraction is performed based on these targets in which the customers opinion has been expressed In recent research of Opinion Mining researchers have applied innovative techniques like target extraction topic modelling in domain independent customer reviews and in Microblogging posts at Facebook and Twitter However these techniques are generally sensitive to extract exact targets and their opinions since they extract the result data in an insignificant way newline