Research Paper
Year: 2020 | Month: April | Volume: 7 | Issue: 4 | Pages: 69-73
Analysis of Performance Cross Validation Method and K-Nearest Neighbor in Classification Data
Sitefanus Hulu1, Poltak Sihombing2, Sutarman2
1Postgraduate Students at Universitas Sumatera Utara, Medan, Indonesia
2Postgraduate Lecturer at Universitas Sumatera Utara, Medan, Indonesia
Corresponding Author: Sitefanus Hulu
ABSTRACT
To produce data classifications that have data accuracy or similarity in proximity of a measurement result to the actual numbers or data, testing can be done based on accuracy with test data parameters and training data specified by Cross Validation. Therefore data accuracy is very influential on the final result of data classification because when data accuracy is inaccurate it will affect the percentage of test data grouping and training data. Whereas in the K-Nearest Neighbor method there is no division of training data and test data. Based on the evaluation results of the Cross Validation algorithm on the effect of the number of K in the K-nearest Neighbor classification data. The data sharing with Cross Validation has better data recognition with a percentage of 100%. The results of the K-NN test results in the classification of data using iris data sets using variation test values 3, 4, 5, 6, 7, 8, 9, have 100% percentage accuracy with 75 true amount of data and 0 incorrect amount of data. Percentage of variation in K K-Nearest Neighbor 3,4,5,6,7,8,9. and variations in the number of K-Fold 1,2,3,4,5,6,7,8,9,10. has a percentage of 100% on K-Fold 4 and 7
Keywords: Classification Data, Cross Validation, K-Nearest Neighbor
[PDF Full Text]