Sentiment Analysis on Online Reviews using Machine Learning and NLTK

Abstract

Advancement in technology paved the way to the new era where machines play great role. Machines make decisions and understand us based on our behaviors and can be trained by data to learn and provide a good accuracy in analyzing business data insights, recommending of goods. The sentiment analysis determines the product effectiveness felt to the customers or public in terms of ratings. The analysis is carried out on modern data science tool and technologies. The machine learning and natural language processing tools and algorithms determines the sentiment classifications either in positive, negative, or neutral class. The accusation of data, through preprocessing and feature extraction in supervised form has been analyzed for the accuracy enhancement on a particular set of algorithms and NLTK tools. The pre architecture was designed and the data sets were then implemented for the effective results and accuracy. In this paper, different machine learning algorithms are applied with natural language kit tools for extraction of features, and it is concluded that logistic regression works best with accuracy of 80% and rest algorithms have less accuracy as compared to logistic regression.