Journal ID : TRKU-11-02-2021-11419
[This article belongs to Volume - 63, Issue - 03]
Total View : 563

Title : Classification of Political Data on Social Media Twitter using Naive Bayes Algorithm

Abstract :

Twitter is social media that can be used to exchange ideas and give opinions. Twitter users can write their opinions on the issue of President Joko Widodo's government. Tweet data or public opinion can be done sentiment analysis method to analyze public opinion. The Naïve Bayes method is used to classify Twitter data to determine sentiment and grouping into positive class and negative class. Furthermore, topic modeling is carried out with the Latent Dirichlet Allocation (LDA) method to determine the topic of discussion in each sentiment group. In the classification process, the value of accuracy depends on the preprocessing stage and relies on the data amount. In train data 80% and test data 20% obtained accuracy 84.58%, recall 85%, precision 85% and F1-Score 85%. At the LDA stage, performance testing with perplexity resulted in a perplexity value of 7.1049 based on the number of iterations of 30 for the positive sentiment group. Furthermore, the perplexity value is 7.3165, with the number of iterations is 60 for the negative sentiment group.

Full article

//