Named entity recognition and tweet sentiment derived from tweet segmentation using hadoop

Abstract

Twitter is well known website famous for micro blogging where millions of users exchanging their opinions and thoughts. The tweets users are sharing has a error sum nature. The information available in tweets is insufficient. Because of character limitation tweets are short in nature many applications like Information Retrieval has problems in information retrieval. Here we are proposing a batch processing framework for tweet fragmentation called TweetSeg. TweetSeg combines information from Confined context with information from Universal context for achieving better results for Named Entity identification. Tweeter is used largely so we want to find public sentiment of tweet by segmenting the tweet into fragments where each fragment can be a named entity, we can find meaningful information from the part and analyzing the sentiments expressed in the tweets by using these fragments in Hadoop framework.