Comparison K-Medoids Algorithm and K-Means Algorithm for Clustering Fish Cooking Menu from Fish Dataset

Abstract

The production of fish-based food processing has become a commodity for restaurants, restaurants, catering and home consumption, but there are still many people who don’t know how fish can be processed in various dishes for their daily needs. To find out how to make fish-based dishes, the researchers provide a solution to cooking any kind of food, starting from the grouping of types of dishes, the basic ingredients that must be prepared, how to cook them, to the address of the cooking link with ingredients from fish. This study aims so that people can cook various menus whose basic ingredients come from fish. This research uses clustering algorithm, k-means and k-medoids. The stages of this research consisted of data collection, data selection, modeling, data training, data testing and evaluation. The object in the study of menu data for various processed fish dishes consisted of 978 datasets of processed fish dishes. The data used for data relating to fish food ingredients with fish food attributes and the number of likes via the website, the fish dataset is sourced from https://ipm.bps.go.id/data/dataset/ikan. From the two algorithms, the best accuracy results are -1.777 for the k-means algorithm, while -1.535 results are obtained for the accuracy of the k-medoids algorithm.