IMPLEMENTASI DATA MINING MENGGUNAKAN METODE DECISION TREE DAN K-NEAREST NEIGHBOR UNTUK MENDETEKSI ANOMALI PADA VOUCHER DISKON PENJUALAN BERBASIS WEB (STUDI KASUS BURGER KING EMERALD)

Abstract

Discount voucher-based promotion is a key marketing strategy at Burger King Emerald to increase sales volume. However, implementation in the field faces challenges in the form of indications of unnatural transactions that have the potential to harm the company. The current detection process is still carried out manually using spreadsheets and subjective observation, making it inefficient and carrying a high risk of human error. This study aims to build a web-based decision support system capable of detecting anomalies in voucher transactions
automatically and accurately. The method applied is data mining with a Combined Model using Decision Tree and K-Nearest Neighbor (K-NN) algorithms. The system development was carried out using the Python programming language with the Flask framework and MySQL database. Based on the test results on 241 sales transaction data in September, the system was able to classify transactions with an accuracy rate of 82.9%. Of the total test data, the system successfully identified 69 transactions (28.6%) as anomalies with the main pattern being disproportionate discount values relative to the total order. Black Box functional test results indicate that all system features are running well. The implementation of this system is expected to assist Burger King Emerald management in minimizing financial losses through faster and more objective early detection of voucher abuse.

Keywords: Data Mining, Anomaly Detection, Decision Tree, K-Nearest Neighbor, Discount Voucher, Website.