In classification, one of the problems that is often encountered is the imbalance class. Unbalanced data occurs when the amount of the data from one class has more or less than the other classes. Classes with unbalanced data will cause the classification results to be skewed towards classes that have more data. Some classifiers are unable to produce maximum accuracy when used on unbalanced data. To overcome the problem of imbalance class, the Synthetic Minority Over-Sampling Technique (SMOTE) method can be used. This method will generate new synthetic data which will be used as training data for the classification process. The method used for classification in this study is the K-Nearest Neighbor (KNN) method. The accuracy value that obtained whe n the data is classified using the KNN method without using SMOTE is 60%. Meanwhile, when the unbalanced data is handled first using SMOTE method and then classified using the KNN method, the accuracy value obtained is 85%. From the test result, the best parameter values were k=1 and N=100.
Copyrights © 2022