<![CDATA[Twitter sentiment analysis is one method of identifying and classifying opinions into positive or negative sentiment in tweets. One of the topics that is being widely discussed on Twitter and has received various opinions for and against is electric buses. All of these opinions are still random and their sentiments have not been classified so sentiment classification needs to be carried out.Naïve Bayes can be used to classify sentiment and is easy to implement. The aim of this research is to classify whether sentiment regarding electric buses leads to positive sentiment or negative sentiment using Naïve Bayes and calculate the accuracy obtained. Several steps were taken, namely data collection, preprocessing, lexicon labeling, word weighting, naïve Bayes classification, and confusion matrix evaluation. The results of this stage from 4 trials of different data sharing ratios showed that the highest sentiment was positive sentiment which reached 77.31% with 22.69% negative sentiment at a data sharing ratio of 6:4 with the evaluation results using the confusion matrix obtaining an accuracy of 74.4%. After naïve Bayes was optimized with hyperparameter tuning, the accuracy increased to 78%. At a data sharing ratio of 9:1, the accuracy obtained after optimization shows a decrease to 71.5%, whereas initially Naïve Bayes obtained an accuracy of 75.6%, this shows that the data split ratio can influence the accuracy obtained by the classification model.]]>
Copyrights © 2024
