<![CDATA[Social media is a place to express or share daily activities. Various new events are often discussed on social media, such as on Twitter. Frequently, the conversations conducted by Twitter users when giving a review or opinion have various emotions, such as anger, sadness, fear, or joy. Emotions are difficult to describe the challenges that occur, sometimes leading to multiple interpretations and misunderstandings leading to debates and reporting to the authorities. So this shows that emotions in reviews and opinions are essential for classification because emotions that come from texts are difficult to understand. In addition, the classification of emotions needs to be done to speed up the identification of emotions. The purpose of this study is to find out which algorithm has optimal performance in the classification of emotions. Machine Learning methods are the Naïve Bayes algorithm, Random Forest, and Support Vector Machines; this is done to determine the dominant algorithm in classifying emotions. The results of the modeling and classification using the Random Forest algorithm obtained a dominant accuracy with an accuracy value of 81.3%, followed by the SVM algorithm with an accuracy value of 76.6% and an accuracy value of 79.1% Naïve Bayes algorithm. In addition, from the speed of time in completing the classification, the Random Forest algorithm has the fastest time of 1.27 seconds]]>
Copyrights © 2022
