EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Segmentasi Pelanggan Dua Tahap Menggunakan K-Means dan DBSCAN Berdasarkan Model RFMA dan Preferensi Desain pada Bisnis Dekorasi Interior

devi sugianti, Arief Soma, Muhammad Faizal, Ari Putra

Jurnal Informatika: Jurnal Pengembangan IT · 22.10.2026 · 11 (3) · 866-878

DOI: 10.30591/jpit.v11i3.35 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Jurnal Informatika: Jurnal Pengembangan IT, online erschienen am 22.10.2026. Verfasst von devi sugianti und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Increasing competition in the interior decoration retail industry requires a deeper understanding of customer behavior to support effective marketing strategies. This study proposes a two-stage customer segmentation framework that combines the RFMA (Recency, Frequency, Monetary, Age) model in the first stage with design preference attributes (product category, price level, room type, and specific items) in the second stage. The study uses transaction data from 500 customers of NAKA Home Decor. In the first stage, K-Means and DBSCAN were compared using Z-score standardized RFMA features. The optimal number of K-Means clusters was determined using the Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Index, while the DBSCAN parameter was selected using a k-distance graph. The results indicate that DBSCAN, which was applied only in the first stage, outperformed K-Means (Silhouette Score 0.477 vs 0.450; Davies-Bouldin Index 0.640 vs 0.801), particularly in identifying customers with extreme transaction values, producing two clusters and eight noise points (1.6%). A sensitivity analysis showed that this result holds under robust scaling of the Monetary variable but not under its logarithmic transformation. In the second stage, each RFMA cluster was further segmented based on design preferences using one-hot encoding. The findings reveal customer preference heterogeneity that cannot be captured by RFMA alone. For example, within the regular-infrequent customer cluster, the Modern-Premium subgroup exhibited an average monetary value approximately 2.4 times higher than the Modern-Midrange subgroup. These findings demonstrate that the proposed two-stage segmentation provides more detailed and actionable customer insights than the conventional single-stage RFM approach.

Zitieren

APA

sugianti, D., Soma, A., Faizal, M., & Putra, A. (2026). Segmentasi Pelanggan Dua Tahap Menggunakan K-Means dan DBSCAN Berdasarkan Model RFMA dan Preferensi Desain pada Bisnis Dekorasi Interior. Jurnal Informatika: Jurnal Pengembangan IT, 11(3), 866-878. https://doi.org/10.30591/jpit.v11i3.35

Vancouver

sugianti D, Soma A, Faizal M, Putra A. Segmentasi Pelanggan Dua Tahap Menggunakan K-Means dan DBSCAN Berdasarkan Model RFMA dan Preferensi Desain pada Bisnis Dekorasi Interior. Jurnal Informatika: Jurnal Pengembangan IT. 2026;11(3):866-878. doi:10.30591/jpit.v11i3.35

BibTeX
@article{sugianti2026,
  author = {sugianti, devi and Soma, Arief and Faizal, Muhammad and Putra, Ari},
  title = {Segmentasi Pelanggan Dua Tahap Menggunakan K-Means dan DBSCAN Berdasarkan Model RFMA dan Preferensi Desain pada Bisnis Dekorasi Interior},
  journal = {Jurnal Informatika: Jurnal Pengembangan IT},
  year = {2026},
  volume = {11},
  number = {3},
  pages = {866--878},
  doi = {10.30591/jpit.v11i3.35}
}
Kostenlos starten →