Market Basket Analysis Menggunakan Algoritme Apriori

Abstract

Abstrak. Market basket analysis is one method in data mining that aims to find products that are frequently purchased together and the association between products in the transaction data in a retail store. This analysis is constructed with 2-step that is looking for a set of products that are often purchased and found an association between products. One of the algorithms used to find the set of frequently purchased products is the a priori algorithms. While the association between products is sought based on two measures, i.e., support and confidence. Support between the two sets of products stating the percentage of the number of sales transactions are two sets of these products together in the overall transaction. While confidence between two sets of products stating the percentage of the number of sales transactions are the two sets of data together in a product throughout the sales transaction that contains one set of products. This paper discusses the market basket analysis that searches the set of products that often are bought using the a priori algorithms. The data used in this paper is the purchase transaction data for 1 year in one of the pastry shop named Extended Bakery in America. The data contains 20.000 transactions with 50 types of products. As a result, the Apriori algorithm obtained 50 frequent 1-itemsets, 26 frequent 2-itemsets, and 4 frequent 3-itemsets as well as products that have to be placed side by side is the Opera Cake, Tart Cherry and Apricot Danish.

 

Abstrak. Market basket analysis merupakan salah satu metode dalam penambangan data (data mining) yang bertujuan untuk menemukan produk-produk yang sering dibeli bersamaan dan asosiasi antar produk dalam data transaksi di toko ritel. Analisis ini dibangun dengan 2 langkah yaitu mencari himpunan produk yang sering dibeli dan menemukan asosiasi antar produk. Salah satu algoritma yang digunakan untuk mencari himpunan produk yang sering dibeli adalah algoritma Apriori. Sedangkan asosiasi antar produk dicari berdasarkan dua ukuran, yaitu support dan confidence. Support antara dua himpunan produk menyatakan persentase banyaknya transaksi penjualan dua himpunan produk tersebut secara bersama-sama dalam keseluruhan transaksi. Sedangkan confidence antara dua himpunan produk menyatakan persentase banyaknya transaksi penjualan dua himpunan produk secara bersama-sama dalam data seluruh transaksi penjualan yang memuat salah satu himpunan produk tersebut. Makalah ini membahas market basket analysis yang pencarian himpunan produk yang sering dibelinya menggunakan algoritma apriori. Data yang digunakan dalam makalah ini adalah data transaksi pembelian selama 1 tahun di salah satu toko kue yang bernama Extended Bakery di Amerika. Data tersebut memuat 1.000 transaksi dengan 50 jenis produk.