Imamah, Sukma Roihatul Nur (2025) Market Basket Analysis Menggunakan Hybrid Association Rule Mining untuk Identifikasi Pola Pembelian Mitra Bisnis pada Transaksi Business-to-Business Berdasarkan Data Penjualan Historis. Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6032232027_Master_Tesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6032232027_Master_Tesis.pdf - Accepted Version Restricted to Repository staff only Download (38MB) | Request a copy |
Abstract
Transaksi Business-to-Business (B2B) memiliki karakteristik khusus berupa volume pembelian yang besar, kontrak jangka panjang, serta pola permintaan yang bervariasi, sehingga produsen dituntut mampu memahami pola pembelian mitra bisnis secara akurat. Perusahaan Lampu XYZ menghadapi tantangan berupa penurunan penjualan dalam tiga tahun terakhir akibat variasi perilaku pembelian mitra bisnis. Untuk mengatasi permasalahan tersebut, penelitian ini mengusulkan penerapan pendekatan Hybrid Association Rule Mining dengan mengombinasikan FP-Growth, Interestingness Measures, dan optimasi menggunakan Non-dominated Sorting Genetic Algorithm-II (NSGA-II). Penelitian ini menggunakan 66.299 transaksi pembelian B2B sebagai dataset. Tahap pertama dilakukan dengan menerapkan FP-Growth dengan ambang batas minimum support sebesar 0,01 dan menghasilkan 4.619 kombinasi frequent itemsets yang kemudian menjadi 18.980 aturan asosiasi awal dengan minimum confidence 0,5. selanjutnya dilakukan penyaringan menggunakan interestingness measures dengan support ≥0,01, confidence ≥0,8, dan lift ≥1 dan menghasilkan 2.966 aturan. Aturan tersebut kemudian dioptimasi menggunakan NSGA-II dengan dua fungsi objektif, yaitu confidence dan lift, untuk mendapatkan aturan yang terbaik. Proses optimasi menghasilkan 7 aturan Pareto optimal yang mewakili pola pembelian paling signifikan dalam transaksi mitra bisnis. Hasil penelitian menunjukkan bahwa pola pembelian didominasi oleh dua kelompok produk, yaitu Lampu Bulb Tipe A (3 W, 5 W, 7 W, 9 W, 11 W, 13 W, 15 W, 18 W, 20 W, 23 W) serta Lampu downlightTipe A, C, dan D dengan variasi watt (5 W, 6 W, 9 W, 11 W, 12 W, 18 W). Produk-produk ini memiliki keterkaitan pembelian yang kuat dan konsisten, sehingga dapat dijadikan dasar dalam perumusan strategi bundling, penguatan promosi produk, serta perencanaan persediaan. Pendekataan hybrid associacion rule mining terbukti mampu menghasilkan aturan yang lebih akurat, terfokus, dan bernilai strategis bagi perusahaan dalam memahami pola pembelian mitra bisnis B2B.
========================================================================================================================
Business-to-Business (B2B) transactions have distinct characteristics, including large purchase volumes, long-term contracts, and highly varied demand patterns. These conditions require manufacturers to accurately understand the purchasing behavior of their business partners. XYZ Lighting Company has experienced declining sales over the past three years due to increasingly diverse purchasing patterns among its distributors. To address this issue, this study proposes the application of a Hybrid Association Rule Mining approach that integrates FP-Growth, Interestingness Measures, and optimization using the Non-dominated Sorting Genetic Algorithm-II (NSGA-II). The study utilizes a dataset consisting of 66,299 B2B purchase transactions. In the first stage, FP-Growth was applied using a minimum support threshold of 0.01, generating 4,619 frequent itemsets, which expanded into 18,980 initial association rules using a minimum confidence of 0.5. These rules were then filtered using Interestingness Measures with support ≥ 0.01, confidence ≥ 0.8, and lift ≥ 1, resulting in 2,966 refined rules. The filtered rules were subsequently optimized using NSGA-II with two objective functions—confidence and lift—to obtain the most representative rules. The optimization process produced seven Pareto-optimal rules that capture the most significant purchasing patterns within the distributor transactions. The results indicate that the purchasing patterns are dominated by two main product groups: Bulb Lamp Type A (3 W, 5 W, 7 W, 9 W, 11 W, 13 W, 15 W, 18 W, 20 W, and 23 W) and Downlight Lamp Types A, C, and D (5 W, 6 W, 9 W, 11 W, 12 W, and 18 W). These products exhibit strong and consistent co-purchasing relationships, making them valuable for developing bundling strategies, strengthening promotional programs, and improving inventory planning. The hybrid association rule mining approach is proven to generate more accurate, focused, and strategically meaningful rules, enabling the company to better understand and respond to the purchasing behavior of its B2B partners.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Business-to-Business (B2B), Hybrid Association Rule Mining, FP-Growth, NSGA-II, Pola Pembelian. |
| Subjects: | H Social Sciences > HB Economic Theory > HB801 Consumer behavior. H Social Sciences > HF Commerce > HF5415.127 Market segmentation. Target marketing H Social Sciences > HF Commerce > HF5438.35 Data Processing (selling) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) |
| Divisions: | Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT) |
| Depositing User: | Sukma Roihatul Nur Imamah |
| Date Deposited: | 28 Jan 2026 04:07 |
| Last Modified: | 28 Jan 2026 04:07 |
| URI: | http://repository.its.ac.id/id/eprint/130220 |
Actions (login required)
 |
View Item |