IMPLEMENTASI SISTEM TRANSAKSI PEMBELIAN BARANG DI MUSTIKA SWALAYAN MENGGUNAKAN POLA ASSOCIATION RULE DAN ALGORITMA APRIORI BERBASIS WEB

Abstract

Every self-service retail business generates a huge amount of transaction data stored in a database, but it is often used only for archiving. However, transaction data can be processed to obtain strategic information that supports business decision-making. This research aims to design and build a web-based application to analyze purchasing transaction patterns at Mustika Supermarket using data mining techniques with the Apriori algorithm. The research process includes several stages: data selection, data cleaning (preprocessing), data transformation, system design, implementation, and testing. The data used are sales transaction records for one year that have been preprocessed to match the algorithm's input format. The application was developed using the PHP programming language with the CodeIgniter framework and a MySQL database, while the software development method applied is the waterfall model. System testing was conducted using a black-box approach to ensure all functions run according to user requirements. The data processing results show the formation of a number of frequent itemsets that meet a minimum support threshold of 3 and a minimum confidence level of 70%. The strongest association rule found is "if a customer buys Indomie, the customer also buys eggs" with a confidence level of 100%. This finding provides important insights for store management to develop marketing strategies, such as product layout arrangements, bundling promotional offers, and more effective inventory planning. Overall, the developed application has proven to be able to assist business owners in exploring consumer purchasing patterns, thereby supporting data-driven decision-making and increasing retail business competitiveness.

Keywords: Apriori Algorithm, Data Mining, Association Rule, Purchase Pattern.