New Approach for Effective Twitter Sentiments Analysis

People can share and express their perspectives on various topics, have discussions with various communities, or post global messages on social networking sites such as Twitter, Facebook, and Google. These sites are quickly gaining popularity because they enable people to do these things. The discipline of analyzing people’s feelings based on the content of tweets on Twitter has seen much progress recently. Finding the most effective sentiment classifier to place tweets in the appropriate categorization is the most challenging aspect of conducting Twitter sentiment analysis. As a result of this study’s findings, a new and improved approach to word representation has been suggested. This approach uses the contribution of sentiment information into the conventional TF-IDF algorithm and generates weighted word vectors. This research provides a comparison between the two methods (TF-IDF and VADER), in addition to the use of machine learning and deep learning algorithms (KNN, RF, LSTM, and ANN), to choose the best algorithms to build a robust model that is capable of accurately analyzing sentiments. This comparison aims to determine which algorithms are the most effective for building such a model. The findings demonstrate that using TF-IDF in conjunction with the RF algorithm results in an accuracy of 98.7%, more significant than the accuracy achieved by other algorithms employing the same TF-IDF or the other VADER technique.