Prediction of Positive and Negative Sentiments for Twitter Data Using Machine Learning

Abstract

Analysis of sentiment is one of the important field in Natural Language Processing (NLP) used to find the emotions using the text. It is an approach to identify the mentality, perspective or feelings of the individual towards anything, administration, film and so on by dissecting the feelings and surveys shared via online media. Different online media networks such as Facebook, Twitter, etc., permit individuals to share their perspectives with others. Twitter become the most well-known online media network that permits clients to share data via the short messages called tweets on a continuous premise. A huge number of individuals associate with one another simultaneously, and a tremendous measure of information is created in seconds. To utilize this information, we build up a Twitter visualization and sentiment analysis system. The dataset used for sentiment analysis is collected from the Twitter website. The pre-processing steps are applied for cleaning the collected dataset by removing hashtags like watchwords. Application Programming Interfaces (APIs) are used to perform sentiment analysis after cleaning the data. This will dissect the estimations as positive and negative for a specific item and administration that helps associations, ideological groups and average folks to comprehend the viability of their endeavours and better dynamic. Our results show that it can handle information progressively, and get envision data consistently.