eprintid: 137870 rev_number: 36 eprint_status: archive userid: 217876 dir: disk0/00/13/78/70 datestamp: 2025-09-10 04:13:37 lastmod: 2025-09-10 04:13:37 status_changed: 2025-09-10 04:13:37 type: thesis metadata_visibility: show creators_name: Salsa Nurahma, - creators_name: Rangga Gelar Guntara, - creators_name: Syti Sarah Maesaroh, - creators_nim: NIM2109502 creators_nim: NIDN0016068805 creators_nim: NIDN0025069005 creators_id: nurahmasalsa1933@upi.edu creators_id: ranggagelar@upi.edu creators_id: sytisarah@upi.edu contributors_type: http://www.loc.gov/loc.terms/relators/THS contributors_type: http://www.loc.gov/loc.terms/relators/THS contributors_name: Rangga Gelar Guntara, - contributors_name: Syti Sarah Maesaroh, - contributors_nidn: NIDN0016068805 contributors_nidn: NIDN0025069005 contributors_id: ranggagelar@upi.edu contributors_id: sytisarah@upi.edu title: PEMANFAATAN ALGORITMA K-MEANS UNTUK KLASTERISASI ULASAN FITUR SPINJAM PADA APLIKASI SHOPEE DI GOOGLE PLAY STORE ispublished: pub subjects: L1 divisions: BIDI_S1_TSK full_text_status: restricted keywords: Spinjam, Analisis Sentimen, K-Means, IndoSBERT, UMAP Spinjam, Sentiment Analysis, K-Means, IndoSBERT, UMAP note: https://scholar.google.com/citations?hl=en&view_op=list_works&gmla=AH8HC4yL9SlGQKPCIyNIMwqBfB-WCVuT8GAmBJVzJNGTQ0hhdKdriiDLuVIpC2HvbtSNNjivXkc2pFf1CrvpeA&user=Xi5fZB0AAAAJ ID SINTA Dosen Pembimbing: Rangga Gelar Guntara: 6738149 Syti Sarah Maesaroh: 6681118 abstract: Perkembangan teknologi informasi dan komunikasi mendorong hadirnya berbagai platform digital yang memfasilitasi interaksi antar pengguna. Salah satu fenomena yang muncul adalah electronic word of mouth (e-WOM) yang menjadi referensi penting bagi calon konsumen dalam menilai kualitas, manfaat, dan risiko suatu layanan. Ulasan pengguna di Google Play Store berperan besar dalam membentuk persepsi dan memengaruhi keputusan, termasuk pada fitur Spinjam di aplikasi Shopee. Namun, ulasan terkait Spinjam masih bercampur dengan ulasan layanan Shopee lainnya sehingga aspek penting mengenai kualitas dan risiko layanan berpotensi tidak teridentifikasi secara jelas. Penelitian ini bertujuan untuk mengelompokkan ulasan pengguna terhadap Spinjam dengan memanfaatkan metode text mining berbasis analisis sentimen menggunakan klasterisasi K-Means. Data penelitian berupa ulasan aplikasi Shopee di Google Play Store pada periode 1 Juni 2024 hingga 1 Juni 2025. Setelah tahap preprocessing, data diubah menjadi representasi vektor menggunakan model embedding IndoSBERT, kemudian direduksi dimensinya dengan Uniform Manifold Approximation and Projection (UMAP) untuk meningkatkan efektivitas klasterisasi. Hasil penelitian menunjukkan nilai silhouette coefficient sebesar 0,521 dengan jumlah klaster optimal k=5, yang menandakan struktur klaster terbentuk dengan cukup baik. Hasil ini membuktikan bahwa kombinasi IndoSBERT, UMAP, dan K-Means dapat meningkatkan performa klasterisasi dalam mengelompokkan ulasan berbasis teks, serta dapat memberikan gambaran yang lebih jelas mengenai pengalaman pengguna terhadap fitur Spinjam. The development of information and communication technology encourages the presence of various digital platforms that facilitate interaction between users. One of the phenomena that has emerged is electronic word of mouth (e-WOM), which is an important reference for potential consumers in assessing the quality, benefits, and risks of a service. User reviews on the Google Play Store play a big role in shaping perceptions and influencing decisions, including the Spinjam feature in the Shopee application. However, reviews related to Spinjam are still mixed with other Shopee service reviews so important aspects regarding service quality and risk are potentially not clearly identified. This study aims to group user reviews of Spinjam by utilizing a text mining method based on sentiment analysis using K-Means clustering. The research data is in the form of reviews of the Shopee application on the Google Play Store in the period from June 1, 2024 to June 1, 2025. After the preprocessing stage, the data is converted into a vector representation using the IndoSBERT embedding model, then the dimensions are reduced with Uniform Manifold Approximation and Projection (UMAP) to increase the effectiveness of clustering. The results showed a silhouette coefficient value of 0.521 with an optimal number of clusters k=5, which indicates that the cluster structure is formed quite well. This result proves that the combination of IndoSBERT, UMAP, and K-Means can improve the performance of clustering in grouping text-based reviews, and can provide a clearer picture of the user experience of the Spinjam feature. date: 2025-08-11 date_type: published institution: Universitas Pendidikan Indonesia department: KODEPRODI61209#Bisnis Digital_S1 thesis_type: other thesis_name: other official_url: https://repository.upi.edu/ related_url_url: https://perpustakaan.upi.edu/ related_url_type: org citation: Salsa Nurahma, - and Rangga Gelar Guntara, - and Syti Sarah Maesaroh, - (2025) PEMANFAATAN ALGORITMA K-MEANS UNTUK KLASTERISASI ULASAN FITUR SPINJAM PADA APLIKASI SHOPEE DI GOOGLE PLAY STORE. S1 thesis, Universitas Pendidikan Indonesia. document_url: http://repository.upi.edu/137870/8/S_BIDI_2109502_Title.pdf document_url: http://repository.upi.edu/137870/2/S_BIDI_2109502_Chapter1.pdf document_url: http://repository.upi.edu/137870/3/S_BIDI_2109502_Chapter2.pdf document_url: http://repository.upi.edu/137870/4/S_BIDI_2109502_Chapter3.pdf document_url: http://repository.upi.edu/137870/5/S_BIDI_2109502_Chapter4.pdf document_url: http://repository.upi.edu/137870/6/S_BIDI_2109502_Chapter5.pdf document_url: http://repository.upi.edu/137870/7/S_BIDI_2109502_Appendix.pdf