Optimasi Metode Naïve Bayes Particle Swarm Optimization Analisis Sentimen Formula E Jakarta Pada Twitter

Abstract

The city of Jakarta plans to hold a Formula E racing event to promote electric cars as the vehicle of the future. The Covid-19 pandemic that hit Jakarta forced the plan to be postponed. The postponement caused a polemic in the community on social media due to the condition of Jakarta being hit by Covid-19 but the Jakarta city government still wants to hold Formula E by paying commitment money to the organizers which is not small. This difference of opinion on social media is used as material for sentiment analysis using the Naive Bayes classification method. The Naive Bayes method, which has a weakness in feature selection, is optimized by applying the Particle Swarm Optimization (PSO) feature selection. The results of the application of PSO optimization on the Naive Bayes method show an increase in performance with an accuracy value of 89.16%, precision 91.10%, recall 86.81% and AUC 0.690.

Keywords: Naive Bayes, Particle Swarm Optimization, Sentiment Analysis, Jakarta E-Prix.