Implementasi Algoritma Apriori Untuk Menentukan Pola Pembelian Produk

Abstract

Flanker Distro or known as Flanker Flag Shop is one of the names of distros engaged in the sale of various types of racing-themed t-shirts, jackets and hats. The problem that occurs in Flanker distro is that sales patterns have not yet been formed, making the owner experience difficulties in determining what products must be provided at a certain time and the stored transaction data is only used as an archive, even though the transaction data can be processed and used as useful information to determine future business strategy. With the large number of existing sales transaction data, it will be difficult if the data is analyzed manually, so to overcome this problem a system is needed to process the data automatically so that it is easy to get sales patterns. The results of this processing will produce transaction information to help determine product sales patterns. The implementation will be made in the form of a web application that uses the UML (Unified Modeling Language) modeling method and the a priori algorithm by providing relationships between items in sales data. In this case, it is a product purchased by a consumer so that a consumer buying pattern will be obtained. The application of the Apriori Algorithm helps in forming possible item combination candidates, then testing whether the combination meets the minimum support and confidence parameters which are the threshold values given by the user. The test was carried out with monthly Flanker distro transaction data from January to July 2022 with a total of 316 sales transaction data. The results of the analysis are obtained after calculating the association rule using the minimum support rule of 5 and a minimum confidence of 30%. So that later it will produce information that can be the basis for Flanker distro owners to determine business strategies and carry out product production in the following months, semesters and years.