Penerapan Association Rule Mining Berbasis Algoritma FP-Max dan Analisis Sensitivitas untuk Rekomendasi Strategi Bundling Produk pada Perusahaan Farmasi

Santoso, Farrel Pradipa Aryasatya Santoso (2026) Penerapan Association Rule Mining Berbasis Algoritma FP-Max dan Analisis Sensitivitas untuk Rekomendasi Strategi Bundling Produk pada Perusahaan Farmasi. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5010221089-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Industri distribusi farmasi sering menghadapi ketimpangan kontribusi penjualan antar Stock Keeping Unit (SKU), di mana sebagian besar produk memiliki frekuensi transaksi sangat rendah dan berpotensi menjadi beban inventori jangka panjang. Strategi product bundling berbasis data, yaitu penggabungan produk fast moving dengan slow moving dalam satu paket, dapat mendorong perputaran SKU yang stagnan tanpa mengorbankan profitabilitas. Penelitian ini merancang strategi tersebut untuk PT Irawan Djaya Agung dengan mengintegrasikan association rule mining dan analisis sensitivitas berbasis Break Even Volume (BEV), menggunakan 358 transaksi multi item dari 14 produk selama periode Januari 2024–Desember 2025. Implementasi FP-Max pada minimum support 2% menghasilkan 21 maksimal kombinasi produk dalam 0,0367 detik dan 3.233 association rules yang lolos filter statistik dengan lift tertinggi 39,78. Dibandingkan FP-Growth, FP-Max mereduksi output sebesar 92,9% (21 berbanding 294 itemset) dan 9 kali lebih cepat. Secara praktis, 21 kombinasi jauh lebih mudah dievaluasi manajemen dibanding 294 yang sebagian besar redundan. Dari rule generation tersebut dikurasi 15 paket bundling dalam empat cluster, kemudian dievaluasi pada tiga skenario diskon (1%, 3%, 5%). Seluruh paket layak pada diskon 1% dan 3% dengan gross profit margin di atas batas minimum perusahaan (14%), sementara pada 5% hanya 5 paket yang bertahan. Diskon 3% ditetapkan sebagai acuan karena menjaga kelayakan semua paket sekaligus memberi penghematan yang cukup terasa bagi pelanggan. Analisis BEV menunjukkan sebagian besar paket hanya butuh 1–7 unit per bulan untuk mencapai titik impas, dengan Paket Treatment Kulit Aktif dan Anti-Jamur & Antiseptik sebagai pilihan implementasi awal yang paling rendah risikonya.
=====================================================================================================================================
The pharmaceutical distribution industry often faces a sales imbalance across Stock Keeping Units (SKU), where most products have low transaction frequency and risk becoming long term inventory burdens. A data driven product bundling strategy, pairing fast moving products with slow moving ones in a single package, is a practical way to push low turnover SKU without sacrificing profitability. This study designs such a strategy for PT Irawan Djaya Agung by integrating association rule mining with Break Even Volume (BEV) based analisis sensitivitas, drawing on 358 multi item transactions from 14 products over January 2024–December 2025. FP-Max at a 2% minimum support produced 21 maximal frequent itemsets in 0.0418 seconds and 3,233 association rules passing the statistical filter with the highest lift at 39.78. Against FP-Growth, FP-Max reduced output by 92.9% (21 versus 294 itemsets) and ran about nine times faster. In practice, 21 maximal combinations are far easier for management to act on than 294 largely redundant ones. The rules were curated into 15 bundling packages across four clusters and tested under three discount scenarios (1%, 3%, 5%). All 15 packages remained viable at 1% and 3% with gross profit margin above the company's 14% minimum; at 5%, only 5 packages held up. The 3% discount was set as the reference because it kept all packages financially sound while offering a saving that customers would actually notice. The BEV analysis found that most packages need just 1–7 units per month to break even, with the Active Skin Treatment and Anti-Fungal & Antiseptic packages standing out as the lowest risk starting points.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Association rule mining, analisis sensitivitas, distribusi farmasi, FP-Growth, FP-Max, maximal frequent itemset, product bundling, Association rule mining, FP-Growth, FP-Max, maximal frequent itemset, pharmaceutical distribution, product bundling, analisis sensitivitas
Subjects:	H Social Sciences > HB Economic Theory > HB801 Consumer behavior. T Technology > TS Manufactures > TS161 Materials management. T Technology > TS Manufactures > TS167 Costs, Industrial
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User:	Farrel Pradipa Aryasatya Santoso
Date Deposited:	30 Jul 2026 12:58
Last Modified:	30 Jul 2026 12:58
URI:	http://repository.its.ac.id/id/eprint/140106

Actions (login required)

View Item