EMOTION DETECTION ON TEXT USING NAVE BAYES ALGORITHM

Abstract

Proper use of emotions at the moment and the right atmosphere can affect the outcome of human activities. Emotions are usually not expressed blatantly by humans (implicitly) and also caused by certain events or situations. The text depicts patterns of events or situations that cause emotions to be revealed with words clearly (explicitly). Text is also the main media in communication using computers (Computer-mediated communications) such as email, blogs and social media. According to the nature of emotions can be classified into two namely positive emotions and negative emotions. Emotion detection is a new and moderately researched field of research, particularly in the linguistic field. In this research the emotion detection will be applied to the status of Facebook using the Naïve Bayes method and the weighted word Count Vector. Before the Count Vector method has been applied, the text will be the test data and the training data in applying the Prapremprosesan which includes folding case, Stopword removal, stemming. The results of this study are the accuracy value of 87,7%.