Sentiment Analysis of Twitter Comments (X) on the Cyanide Coffee Case Using Comparison of Naïve Bayes and K-NN Method Results

Abstract

Twitter is a social media platform that is often used by the general public to communicate and gather information. Through this social media, users can express various opinions or comments about a particular issue such as the cyanide coffee case. The methods used in this research are Naïve Bayes and K-NN methods and the purpose of this research is to compare the two methods in sentiment analysis of the cyanide coffee case. In testing this sentiment analysis using 3 stages of scenarios, where the 60:40 ratio is used for scenario number 1, the 70:30 ratio is used for scenario 2 and the 80:20 ratio is used for scenario 3. From the test results that have been carried out, it can be seen that testing the Naïve Bayes method in scenario 3 is superior to other scenarios with 87.76% accuracy, 92.80% precision and 67.09% recall. Meanwhile, the K-NN method in scenario 3 is superior to other scenarios with an accuracy of 80.60%, precision of 89.87% and recall of 61.72%.