Ref | Contribution | Dataset | Algorithms | Best Accuracy |
This Work | An examination at how lexicon-based and deep learning-based approaches to Twitter sentiment analysis vary | Sentiment140 dataset containing 1.6 million tweets. | Lexicon-based approach and Long Short-Term Memory (LSTM) in deep learning | 98% (LSTM) |
[11] | Twitter sentiment analysis in Hausa with the use of machine learning and lexicon-based techniques | BBC Hausa Twitter API. | Logistic Regression, Multinomial Naive Bayes (MNB) with Count Vectorizer and TF-IDF | 86% (LR) |
[12] | Real-time sentiment analysis using SVM, K-SVM, and Multinomial Naive Bayes. | Twitter data and contested territory news websites. | Support Vector Machine (SVM), Kernel SVM (K-SVM), Multinomial Naive Bayes | 80% (K-SVM) |
[13] | Comparative Research of Lexicon and Machine Learning Techniques for Sentiment Analysis | TF-IDF | Logistic regression, support vector machine (SVM), logistic regression, AFINN, and Vader lexicon | 96.3% (SVM) |
[15] | Analysis of Twitter users’ COVID-19 attitudes using machine learning. | COVID-19 tweets from Kaggle. | Naive Bayes, Random Forest, Support Vector Machine, Neural Network | 79.37% (Random Forest) |
[17] | Using deep learning and lexicons, Indian tweets on COVID-19 and vaccination are analyzed. | COVID-19 and vaccination tweets from India. | Recurrent Neural Network (RNN) with Bi-LSTM and GRU techniques | 93.03% (GRU) |