Perbandingan Metode Naïve Bayes dan KNN dalam Penerapan Data Mining untuk Klasifikasi Menu Potensial ( Studi Kasus : Ateku Kopi Medan )

Abstract

The growth of the culinary business, particularly in the café sector, faces intense competition, necessitating effective marketing strategies. The promotional systems employed by business operators to increase sales are sometimes inefficient. This research aims to develop a web-based system using Naïve Bayes and K-Nearest Neighbor (KNN) methods to classify potential menus at Ateku Kopi Medan. The methodology includes literature review, sales data collection from January 2024 to April 2024, system implementation using PHP, and system testing with evaluation using a confusion matrix. A dataset of 428 sales transactions is used, divided into training data (80%) and testing data (20%). The system considers criteria such as price, number sold, and whether a discount is present. The results of the study show that the developed system can provide accurate and effective results in finding potential menus, reducing the risk of errors and increasing efficiency and can identify menus with high, medium, or low potential, which is useful for increasing sales and efficiency of food stock with an accuracy of 90.58% for the Naïve Bayes method and 88.23% for K-Nearest Neighbor (KNN).