Journal ID : TRKU-11-02-2021-11420
[This article belongs to Volume - 63, Issue - 03]
Total View : 564

Title : Classification of Political Data on Social Media with Support Vector Machine Algorithm

Abstract :

This study will classify sentiment and analysis of general topics against President Joko Widodo taken through social media Twitter. The influence and benefits of sentiment are so significant that the study of analytical sentiment is increasing. This research has several stages, including data collection, data cleaning, data extraction, data classification using the Support Vector Machine method, and topic analysis using the Latent Dirichlet Allocation method. This study's results took data on Twitter taken from September 23, 2018, to December 15, 2020, as many as 1200 pieces of data in the form of 600 negative data and 600 positive data. In conclusion, the topic analysis system and sentiment classification performed the classification results using the Support Vector Machine method in this research, which obtained 86.39% accuracy by 85.75%, recall by 85.86%, and F1 score 85.79%—as for topic analysis using Latent Dirichlet Allocation got a perplexity value of 7.1049 on positive sentiment and 7.3165 on negative sentiment.

Full article

//