Metode Naive Bayes Classifier untuk Analisis Sentimen Studi Kasus Mega Wisata Coastarina Batam

Abstract

Tourism is one of the destinations for tourists to visit a city. Usually, to visit a tourist spot, individuals check the ratings of the place through Google reviews. In addition to ratings, some people also read reviews to assess the quality of the tourist spot. However, if the reviews given do not match the rating, there is a need for sentiment analysis to identify positive and negative reviews. This study takes a case study from Batam city, namely Mega Wisata Coastarina Batam. The method used in this research is Naive Bayes. Naive Bayes is chosen for its popularity in classification methods with good accuracy. Based on the research results, there are more positive sentiments than negative sentiments, namely 200 positive sentiments and 141 negative sentiments. The testing results of the research show the performance of sentiment analysis using the Naive Bayes method, with an F1-score accuracy of 69.49% and a margin of error of +-7.67%. The precision is 71.00% in positive predictions and 67.39% in negative predictions. The recall value is 75.53% for positive predictions and 62.09% for negative predictions. The F1-score results are 86.06% for positive predictions and 64.63% for negative predictions.