Klasterisasi Pola Pembelian Konsumen Menggunakan Improved K-Medoids Berbasis Crow Search Algorithm dan Fp-Growth (Studi Kasus: Gerai Makanan Cepat Saji Surabaya)

Azzahra, Melinda Putri (2026) Klasterisasi Pola Pembelian Konsumen Menggunakan Improved K-Medoids Berbasis Crow Search Algorithm dan Fp-Growth (Studi Kasus: Gerai Makanan Cepat Saji Surabaya). Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (cover) 22083010035.-cover.pdf Download (2MB)
	Text (bab 1) 22083010035.-bab 1.pdf Download (250kB)
	Text (bab 2) 22083010035.-bab 2.pdf Restricted to Repository staff only until 15 September 2028. Download (1MB)
	Text (bab 3) 22083010035.-bab 3.pdf Restricted to Repository staff only until 15 September 2028. Download (530kB)
	Text (bab 4) 22083010035.-bab 4.pdf Restricted to Repository staff only until 15 September 2028. Download (1MB)
	Text (bab 5) 22083010035.-bab 5.pdf Download (190kB)
	Text (daftar pustaka) 22083010035.-daftarpustaka.pdf Download (179kB)
	Text (lampiran) 22083010035.-lampiran.pdf Restricted to Repository staff only until 15 September 2028. Download (205kB)

Abstract

The quick service restaurant (QSR) industry in Indonesia continues to grow in line with changes in people’s lifestyles that prioritize speed and convenience. Increasing competition among QSR outlets has encouraged the need to utilize transaction data to better understand transaction characteristics and purchasing patterns. This study applies an Improved K-Medoids method based on the Crow Search Algorithm (CSA) to cluster transactions, followed by the FP-Growth algorithm to analyze purchasing patterns within each cluster. The research data were obtained from the point-of-sale (POS) system of a Hisana Fried Chicken outlet in Surabaya for the period from January 1 to March 28, 2025. The dataset consisted of 18,814 item level records, which were aggregated into 12,841 unique transactions. The preprocessing stages included attribute selection, data quality checking and correction, data type transformation, aggregation, Min-Max Scaling normalization, and exploratory data analysis. The clustering results identified four optimal clusters with a Silhouette Score of 0.5594, a Davies-Bouldin Index (DBI) of 0.7703, and a fitness value of 15.1700. Compared with conventional K-Medoids, the CSA-based method produced better clustering quality based on these three metrics. Sensitivity analysis identified 1,403 transactions as outliers and showed that their presence affected the cluster structure. FP-Growth without clustering produced 35 frequent itemsets but no association rules, whereas after clustering, association rules were found in Clusters 0, 2, and 3. The best rule was found in Cluster 0, where the purchase of Saos BBQ and Sambal Samyang was followed by the purchase of Ayam Bekakak, with 85% confidence and a lift value of 22.8. The research results were presented through a graphical user interface (GUI) to help outlet managers understand the analysis results as a basis for considering datadriven marketing strategies.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Saputra, Wahyu Syaifullah Jauharis	NIDN0725088601	wahyu.s.j.saputra.if@upnjatim.ac.id
Thesis advisor	Nasrudin, Muhammad	NUPTK4241774675130323	nasrudin.fasilkom@upnjatim.ac.id

Subjects:

H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA76.6 Computer Programming

Divisions:

Faculty of Computer Science > Departemen of Data Science

Depositing User:

Melinda Melinda Putri Azzahra

Date Deposited:

16 Sep 2026 06:30

Last Modified:

16 Sep 2026 09:15

URI:

https://repository.upnjatim.ac.id/id/eprint/60235