A Campaign Mining in Social Media using Improved K-Means: The Perspective of Candidate President's Mission

• Published in: 2022 5th International Seminar on Research of Information Technology and Intelligent Systems (ISRITI) pp. 230 - 235
• Main Authors: Zaman, Badrus, Putra, Cendra Devayana, Justitia, Army
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08-12-2022
• Subjects: Blogs; campaign; Clustering algorithms; Information technology; Intelligent systems; k-means; mission; presidential election; Seminars; Social networking (online); Space missions; tweet
• DOI: 10.1109/ISRITI56927.2022.10052929

The presidential candidate socializes the vision and mission to persuade supporters and voters as much as possible during the campaign. Since 2010, presidential campaigns in Indonesia have extensively used social media networks as a standard form of presidential communication. However, the presidential candidate's campaign duration has not been maximized to convey the vision and mission to prospective voters. This study aims to identify each candidate's focus based on tweet analysis using Improved K-Means and a second-level normalization algorithm. The data consisted of 2,786 tweets from January to April 2019, obtained from two sources: 1,485 tweets from candidate#1 and 1,301 tweets from candidate#2. The results show that there are 16 clusters created from candidate#1 and 14 from others. Most candidate#1 aligns with its mission. Consequently, our study could clarify the presidential candidate's mission focus. The more frequently the mission is discussed on social media, the more notable it is as the President's primary work program. The first candidate's tweets mainly mentioned the fourth mission, and the second one was about the first mission. Furthermore, there is a link between the frequency of tweets from the presidential candidate's mission and the theme of the presidential debate.