Implementation of Apriori Algorithm for Data Mining on Sales Transaction Data

Abstract

Angkasa Mart Store is currently experiencing a decline in sales for specific products, leading to the implementation of the Apriori Algorithm to create bundled offerings that combine less popular items with top-selling products, aiming to revitalize sales and promote the underperforming inventory. Following the CRISP-DM methodology, the study analyzes sales transaction data from June to July 2022, covering 65,892 purchased items, to extract ten critical association rules essential for devising bundled packages. The study's findings propose two strategies for package composition: first, the development of bundled packages comprising strongly related products, identified through comprehensive data analysis and positively received by the store; second, the introduction of 21 bundled packages consisting of products with a relatively weaker relationship, effectively expanding consumer choices and encouraging additional purchases within the store. By implementing the Apriori Algorithm and adhering to the CRISP-DM methodology, this study effectively formulates bundled product packages for Angkasa Mart Store, addressing the challenge of declining sales and contributing to an overall improvement in business performance and customer satisfaction.

Keywords

• apriori algorithm, association rule, bundle, data mining