The occurrence of chemical composition deviations or mechanical properties in steel production causes the clasification of steels based on standards can't be defined. The deviation that occurs is the deviation from the maximum limit of the chemical composition and the minimum limit of the mechanical properties of steel. This is the background of researchers to create a system using the Fuzzy k-Nearest Neighbor method. The Fuzzy k-Nearest Neighbor (Fuzzy K-NN) method used for classifying steel standards based on the chemical composition of the steel produced. The data used for this study is data steel products with the specifications of the steel composition, the mechanical properties of the steel and the classification of standard of steel produced. The steps performed are data normalization, Fuzzy k-Nearest Neighbor, calculate Euclidean distance, take the shortest distance k, calculate the membership value of each class and determine the target class. The highest accuracy resulted by testing k values using k-fold cross validation is 74,44% with k value equal to 74,44 and total of training data is 267 data.
Copyrights © 2018