Particle swarm optimization for feature selection in sentiment analysis on the application of digital payments OVO using the algorithm of Naive Bayes

• Main Authors: Hyuningtyas, Ratih Yulia, Sari, Retno, Yusnaeni, Wina
• Format: Conference Proceeding
• Language: English
• Subjects: Accuracy; Algorithms; Data mining; Feature selection; Optimization; Particle swarm optimization; Sentiment analysis
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0129011

The Benefits of the digital payment application is to make the transaction easy without spending a certain amount of money. There are a lot of digital payment applications available in the play store that can be used, but there are still many people who do not use the digital payment application. In deciding to use the digital payment app, people usually see the reviews first to find out what features are provided and how great they give convenience during the transaction. Reviews of digital payment application can become data mining that contains textual information related to sentiment analysis. Sentiment analysis is to analyze a review categorized as the positive or negative. In this study, the researchers tried to analyze a review of a digital payment application using the Naive Bayes algorithm and particle swarm optimization. The purpose of this research is to find the highest accuracy of each experiment, the data used in the trial are classified into the class of positive and negative. The results of this study obtained accuracy with the Naive Bayes algorithm by 82,00%. The accuracy is trying to utilize feature selection to obtain more accurate classification results. Feature selection used in this study, namely particle swarm optimization, consists of a set of particles that is called the population. The experiment was conducted again with the Naive Bayes algorithm with particle swarm optimization that produces the accuracy of 89,00%. The results of the research is an increase in the level of accuracy by using PSO.