Twitter based sentiment analysis to predict public emotions using machine learning algorithms

Abstract

Twitter is a prominent social media platform where users may send and receive messages known as "tweets." Individuals can use this to communicate their opinions or opinions regarding a variety of topics. Sentiment analysis has been performed on such tweets by a variety of parties, including consumers and advertisers, in order to get insights about goods or conduct market research. Additionally, recent advances in machine learning techniques have enhanced the exactitude of sentiment analysis forecasts. In this work, sentiment analysis on "tweets" was performed utilizing a variety of machine learning approaches. It attempts to grade the tweet's polarity as either positive or negative. If a tweet good and negative rudiments, the overall mood ought to be used to classify it. In this research work, Kaggle dataset was used and that had been crawled and categorized as positive or negative. Emoticons, usernames, and hash tags are included in the data, which must be processed and transformed into a standard format. The suggested research project must also extract relevant aspects from the text, just as unigrams and bigrams, that are two different ways to express a "tweet." Ensembling is a type of meta learning algorithm methodology in which researchers mix many classifiers to increase prediction accuracy. Finally, the study shows that Deep Learning approaches outperform other methods.