Algoritma K-Nearest Neighbor dengan Euclidean Distance dan Manhattan Distance untuk Klasifikasi Transportasi Bus
Rozzi Kesuma Dinata, Hafizal Akbar, Novia Hasdyna
Abstract
K-Nearest Neighbor is a data mining algorithm that can be used to classify data. K-Nearest Neighbor works based on the closest distance. This research using the Euclidean and Manhattan distances to calculate the distance of Lhokseumawe-Medan bus transportation. Data that used in this research was obtained from the Organisasi Angkutan Darat Kota Lhokseumawe. The results of the test with k = 3 has obtained the percentage of 44.94% for Precision, 37.06% Recall, and 81.96% Accuracy for the performance of K-NN with Euclidean Distance. Whereas by using Manhattan Distance the result obtained was 45.49% for Precision, 36.39% Recall, and 84.00% Accuracy. The result shown that Manhattan Distance obtained the highest accuracy, with the difference of 2.04% higher than Euclidean Distance. It indicates that Manhattan Distance is more accurate than Euclidean Distance to classify the bus transportation.