Implementation of the K-Nearst Neighbor (k-NN) Algorithm in Classification of Angora and Country Cats

Abstract

There are so many types of mixed cats from various cat breeds, so many people find it difficult to identify and classify them. Therefore, we need a method that can classify the type of cat breeds. In this study the authors used the K-Nearest Neighbor (k-NN) algorithm to make it easier to recognize and classify cat breeds based on certain characteristics. The author took samples of 2 types of cat races, namely the Anggora race and the Kampung race. The implementation stage is to determine the euclidean distance and sort it and then determine the value of K to find the nearest neighbor. In testing, the authors used 50 training data and 50 test data with 6 attributes used, namely body shape, nose width, nose height, food type, hair type and hair length. The results of the classification of cat breeds using the k-NN method obtained an accuracy rate of 94% and an error rate of 6%.