Controversial Analysis:- Sentimental Analysis of Twitter Data

Samarth Jaykar Shetty, Badal Rakesh Thosani, Lenherd Deon Olivera, Supriya Kamoji
2017 International Journal of Advanced Research in Computer Science and Software Engineering  
Controversial analysis deals with identifying and classifying opinions or sentiments expressed in source text. We present a novel approach for naturally ordering the sentiments of Twitter messages. These messages are delegated positive or negative or neutral concerning a query term. This is valuable for buyers who need to examine the notion of items before buy, or organizations that need to screen the general public sentiment of their brands. Past research on characterizing opinion of messages
more » ... pinion of messages on microblogging administrations like Twitter have attempted to tackle this issue however have disregarded neutral tweets which prompts to wrong feeling characterization and we have attempted to take care of this issue in this project. We show the consequences of machine learning algorithm for classifying the sentiment of Twitter messages utilizing a novel feature vector. Our training data comprises of openly accessible twitter messages acquired through mechanized means. We demonstrate that machine learning algorithm (Naive Bayes and SVM) can accomplish aggressive exactness when prepared utilizing our feature vector and the freely accessible dataset. This report also describes the pre-processing steps of the dataset required to accomplish high accuracy. The primary commitment of this project is the novel feature vector of weighted unigrams used to prepare the machine learning classifiers.
doi:10.23956/ijarcsse/v7i4/0124 fatcat:admxhugvefd5pn3zbhp6rxtvve