Analisis Pola Transaksi Bernilai Tinggi Dengan Algoritma Closed Frequent And High Utility Itemset Mining (Closed-FHUIM)

Sulanjari, Kinana Syah (2025) Analisis Pola Transaksi Bernilai Tinggi Dengan Algoritma Closed Frequent And High Utility Itemset Mining (Closed-FHUIM). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032231187-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Bisnis ritel menghadapi tantangan dalam mengelola inventaris akibat fluktuasi permintaan yang tidak menentu dan kurangnya pemahaman terhadap pola pembelian pelanggan. Kondisi ini dialami pula oleh Kopkar Sampoerna, koperasi yang melayani konsumen ritel dan mitra Business-to-Business (B2B), khususnya dalam mengidentifikasi pola pembelian yang bernilai tinggi dan berfrekuensi tinggi. Penelitian ini bertujuan untuk mengidentifikasi pola pembelian tertutup yang bernilai tinggi dan sering terjadi menggunakan algoritma Closed Frequent and High Utility Itemset Mining (FCHUIM) dengan optimasi OSR, OWL, dan MSU. Data transaksi selama dua tahun digunakan untuk membangun model FCHUIM dan dibandingkan dengan data tren terbaru tiga bulan terakhir untuk validasi temporal.
Hasil eksperimen menunjukkan bahwa algoritma FCHUIM berhasil menemukan pola-pola itemset yang tidak hanya sering terjadi tetapi juga memberikan kontribusi besar terhadap nilai utility. Evaluasi dilakukan menggunakan metrik support, utility, dan panjang itemset. Perbandingan antara hasil algoritma FCHUIM standar dan versi optimasi menunjukkan adanya efisiensi dalam jumlah pola yang dihasilkan, waktu eksekusi, dan penggunaan memori. Selain itu, dilakukan juga analisis what-if terhadap perubahan demand ±10% dan simulasi bundling strategi untuk mendukung pengambilan keputusan pemasaran dan manajemen stok. Validasi temporal menunjukkan bahwa 44,19% pola lama tetap relevan dalam tren terbaru.
Penelitian ini memberikan kontribusi dalam pengembangan strategi bundling, penghematan potensi stok, dan pengambilan keputusan berbasis data dalam
manajemen ritel. Pendekatan FCHUIM terbukti mampu menghasilkan pola yang lebih ringkas, relevan, dan bernilai dalam konteks manajemen inventaris modern.

=======================================================================================================================================

Retail businesses face challenges in managing inventory due to unpredictable demand fluctuations and a lack of insight into customer purchasing patterns. This issue is also experienced by Kopkar Sampoerna, a cooperative serving both retail customers and Business-to-Business (B2B) partners,
particularly in identifying high-value and frequently occurring purchase patterns. This study aims to identify closed, high-utility, and frequent purchase patterns using the Closed Frequent and High Utility Itemset Mining (FCHUIM) algorithm with OSR, OWL, and MSU optimizations. Two years of transaction data were used to build the FCHUIM model, and validation was conducted using the most recent three-month transaction trends.
The experimental results show that the FCHUIM algorithm successfully discovered itemset patterns that are not only frequent but also significantly contribute to utility value. The evaluation was carried out using support, utility, and itemset length metrics. A comparison between the standard and optimized FCHUIM algorithms indicates increased efficiency in terms of the number of patterns generated, execution time, and memory usage. Additionally, a what-if analysis was conducted to simulate ±10% changes in demand, along with bundling strategy simulations to support marketing and stock management decisions. Temporal validation revealed that 44.19% of the previous patterns remain relevant in current transaction trends.
This research contributes to the development of bundling strategies, potential stock optimization, and data-driven decision-making in retail inventory management. The FCHUIM approach has proven effective in producing more concise, relevant, and valuable patterns in the context of modern inventory management.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Frequent Itemset Mining (FIM), High Utility Itemset Mining (HUIM), Closed Frequent and High Utility Itemset Miner (Closed-FHUIM), Pola pembelian pelanggan, Frequent Itemset Mining (FIM), High Utility Itemset Mining (HUIM), Closed Frequent and High Utility Itemset Miner (Closed-FHUIM), Consumer Purchase Pattern.
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.62 Decision support systems T Technology > T Technology (General) > T59.7 Human-machine systems.
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Kinana Syah Sulanjari
Date Deposited:	30 Jul 2025 15:22
Last Modified:	30 Jul 2025 15:22
URI:	http://repository.its.ac.id/id/eprint/123951

Actions (login required)

View Item