SEGMENTASI RESPON PENGGUNA TWITTER/X TERHADAP PENGHAPUSAN TIKTOK SHOP DI INDONESIA MENGGUNAKAN METODE DBSCAN DAN K-MEANS

RAMADHANI, TIARA (2024) SEGMENTASI RESPON PENGGUNA TWITTER/X TERHADAP PENGHAPUSAN TIKTOK SHOP DI INDONESIA MENGGUNAKAN METODE DBSCAN DAN K-MEANS. S1 thesis, Universitas Mercu Buana-Menteng.

	Text (Cover) 41520110027-TIARARAMADHANI-01 Cover - TIARA RAMADHANI.pdf Download (448kB)
	Text (Abstrak) 41520110027-TIARARAMADHANI-02 Abstrak - TIARA RAMADHANI.pdf Download (192kB)
	Text (Bab 1) 41520110027-TIARARAMADHANI-03 Bab 1 - TIARA RAMADHANI.pdf Restricted to Registered users only Download (205kB)
	Text (Bab 2) 41520110027-TIARARAMADHANI-04 Bab 2 - TIARA RAMADHANI.pdf Restricted to Registered users only Download (534kB)
	Text (Bab 3) 41520110027-TIARARAMADHANI-05 Bab 3 - TIARA RAMADHANI.pdf Restricted to Registered users only Download (310kB)
	Text (Bab 4) 41520110027-TIARARAMADHANI-06 Bab 4 - TIARA RAMADHANI.pdf Restricted to Registered users only Download (420kB)
	Text (Bab 5) 41520110027-TIARARAMADHANI-07 Bab 5 - TIARA RAMADHANI.pdf Restricted to Registered users only Download (193kB)
	Text (Daftar Pustaka) 41520110027-TIARARAMADHANI-08 Daftar Pustaka - TIARA RAMADHANI.pdf Restricted to Registered users only Download (236kB)
	Text (Lampiran) 41520110027-TIARARAMADHANI-09 Lampiran - TIARA RAMADHANI.pdf Restricted to Registered users only Download (578kB)
	Text (Lembar Keabsahan) 41520110027-TIARARAMADHANI-10 Hasil Scan Formulir Penyerahan Keabsahan dan Persetujuan Publikasi Tugas Akhir - TIARA RAMADHANI.pdf Restricted to Repository staff only Download (85kB)

Abstract

Media sosial, khususnya Twitter/X, telah menjadi wadah utama bagi masyarakat untuk menyuarakan pendapat dan reaksi terhadap isu-isu kontroversial, termasuk penghapusan TikTok Shop di Indonesia. Fenomena tersebut menimbulkan berbagai kontroversi dan dampak signifikan terhadap ekonomi, terutama UMKM lokal. Penelitian ini menggunakan metode text clustering, yaitu Density-Based Spatial Clustering of Applications with Noise (DBSCAN) dan K-Means, untuk mengelompokkan pendapat masyarakat yang terungkap dalam cuitan-cuitan di Twitter/X terkait penghapusan TikTok Shop. Data penelitian diperoleh melalui crawling tweet yang berkaitan dengan isu tersebut. Data yang berhasil di crawling sejumlah 3.000 tweet sejak tanggal 4 September 2023 sampai 1 Desember 2023, lalu dilakukan proses seleksi data sehingga menjadi 1.185 tweet. Segmentasi respon pengguna menggunakan metode K-means dan DBSCAN menjadi fokus utama, dengan tujuan mendapatkan pemahaman yang lebih dalam mengenai beragam pandangan dan perasaan masyarakat terhadap kebijakan pemerintah. Selain itu, penelitian ini bertujuan untuk memberikan informasi terkait metode yang lebih unggul antara K-Means dengan DBSCAN dalam segmentasi respon pengguna di Twitter/X. Sejumlah penelitian sebelumnya menunjukkan bahwa metode DBSCAN sering kali lebih unggul dalam mengelompokkan data teks dibandingkan K-Means, namun ada pula yang menunjukkan metode K-Means lebih unggul dibanding DBSCAN dalam pengelompokkan data teks. Hasil penelitian ini dihasilkan bahwa metode K-Means lebih unggul dari DBSCAN dalam mengelompokkan data teks yang memiliki variasi tinggi seperti data tweet. Didapat 3 segmen dengan focus permasalahan yang berbeda-beda dari hasil pengelompokkan dengan K-Means, dan dari masing-masing segmen tersebut cenderung memiliki sentiment negative. Social media, particularly Twitter/X, has become a major platform for people to voice their opinions and reactions to controversial issues, including the removal of TikTok Shop in Indonesia. This phenomenon has sparked various controversies and significant economic impacts, especially on local SMEs. This research utilizes text clustering methods, namely Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and K-Means, to group public opinions expressed in tweets on Twitter/X regarding the removal of TikTok Shop. The research data was obtained by crawling tweets related to the issue. A total of 3,000 tweets were crawled from September 4, 2023, to December 1, 2023, and after data selection, 1,185 tweets were used. The primary focus is on segmenting user responses using K-means and DBSCAN methods to gain a deeper understanding of the diverse views and feelings of the public towards government policy. Additionally, this study aims to provide information on which method, K-Means or DBSCAN, is superior in segmenting user responses on Twitter/X. Previous research has shown that DBSCAN often outperforms K-Means in text clustering, while others have indicated that K-Means is superior to DBSCAN in text data clustering. The results of this study show that the K-Means method outperforms DBSCAN in clustering text data with high variability such as tweet data. Three segments with different focal issues were identified using K-Means clustering, and each of these segments tends to have a negative sentiment.

Actions (login required)

View Item