COMPARISONAL ANALYSIS OF EUCLIDEAN, CANBERRA, AND CHEBECHEV DISTANCE MODELS ON KNN METHOD ON STUDENTS' VALUE

Ragil Satya Adi W (Faculty of Science and Technology, Program Study Computer System, University Pembangunan Panca Budi Medan, North Sumatra, Indonesia)
Eko Hariyanto (Faculty of Science and Technology, Program Study Computer System, University Pembangunan Panca Budi Medan, North Sumatra, Indonesia)
Zulham Sitorus (Faculty of Science and Technology, Program Study Computer System, University Pembangunan Panca Budi Medan, North Sumatra, Indonesia)

INFOKUM
Vol. 10 No. 03 (2022): August, Data Mining, Image Processing, and artificial intelligence

KNN has a significant influence on nonparametric methods in the form of classification, but the level of performance generally depends on the equilibrium point of the variable that is correlated with the far point. The distance between readings from the specified limit of the standard deviation value. KNN method. One of the instance-based learning groups is the K-Nearest Neighbor (KNN) method. Group search performed by KNN on new data objects or k objects in the test that is closest to the test data. KNN helps classify objects based on training data that is close to the object being tested. This study concluded that the Canberra Distance model produced the highest accuracy of 87.50% with an error value of 12.50% on the K-Nearest Neighbor algorithm.