Penggunaan Metode Naïve Bayes Classifier pada Analisis Sentimen Facebook Berbahasa Indonesia

• Main Authors: Putu Sri Merta Suryani, Linawati Linawati, Komang Oka Saputra
• Format: Article
• Language: English
• Published: Universitas Udayana 2019-05-01
• Series: Majalah Ilmiah Teknologi Elektro
• Online Access: https://ojs.unud.ac.id/index.php/JTE/article/view/47671

Sentiment analysis is a field that is currently in great demand by various groups. Sentiment analysis can be done using documents and opinions from social media. One social media that is usually used as a means of opinion is Facebook social media. Before a text is classified, it is necessary to do POS Tagging which is the word labeling stage where the purpose is to determine the words which include opinions and non opinions. For labeling words can use the Hidden Markov Model or Rule Based. The method commonly used in sentiment analysis is the Naïve Bayes Classifier method. This method simply classifies probabilities. Naïve Bayes Classifier can be used to classify opinions into positive and negative opinions. In addition, this method uses training data in the classification process. The classification produced from the Naïve Bayes Classifier method is quite good. To test the accuracy of the system in classifying opinions, the classification results are tested. From the test results obtained an average accuracy of 87.1%. The more training data that is similar to testing data, the better the classification results.