Sentiment Analysis on "Homecoming Tradition Restriction" Policy on Twitter

• Published in: 2020 3rd International Conference on Computer and Informatics Engineering (IC2IE) pp. 75 - 79
• Main Authors: Suroso, Heru, Budi, Indra, Santoso, Aris Budi, Putra, Prabu Kresna
• Format: Conference Proceeding
• Language: English
• Published: IEEE 15.09.2020
• Subjects: Blogs; Classification algorithms; covid-19; Data models; Labeling; public policy; Sentiment analysis; Social networking (online); Support vector machines
• ISBN: 172818245X; 9781728182452
• DOI: 10.1109/IC2IE50715.2020.9274609

The "Homecoming Tradition Restriction" was one of the government's policies to terminate and limit the spread of the Covid-19 Virus. Apart from being a media for socializing government policies, Twitter can be utilized by the public to convey responses, opinions, and criticisms towards government policies. This study aims were to determine public sentiment towards the "Homecoming Tradition Restriction" policy. This study uses a data mining approach to classify public sentiments delivered via Twitter. Sentiment classification models are built using two algorithms, Support Vector Machine (SVM) and Naïve Bayes. Naïve Bayes produces the highest performance measurement with a recall of 80% and an F-measure of 71.32%. This study shows that the majority of people support this government policy as indicated by the majority of sentiments that have been collected is positive, and also backed by the fact that the total number of homecoming vehicles during Eid Holiday was decreased by 62% from the previous year. This shows that social media data is relevant enough to be used in the assessment of public responses to government policies.