Automated Sentiment Analysis of Text Data with NLTK

Abstract

At present, most of researches in natural language processing primarily are focused on deep learning way to solve accuracy problems. This paper discusses a branch of natural language processing, sentiment analysis. In order to implement the function of sentiment analysis efficiently, we built an algorithm by calling the NLTK library firstly. As a result, this method can roughly obtain the emotional scores of different sentences and compare them with the scores given by human being who create these sentences. The results are presented by correlation calculation and images such as boxplot. The analysis of the images demonstrates the deficiency of the method. In addition, for the results, we found the direction of learning to solve the accuracy problems and made reasonable arrangements and planning.