<![CDATA[Counterfeiting is currently on the rise in Indonesia. Based on Law No. 7 of 2011 concerning counterfeit rupiah currency, counterfeit currency is defined as any object whose material, size, color, image, and/or design resembles the rupiah that is made, formed, printed, duplicated, circulated, or used as a means of payment illegally. The objectives of this study are: To obtain a paper money classification model by applying the KNN and Gabor Filter algorithms. To improve the evaluation results of the paper money classification model by applying the KNN and Gabor Filter algorithms with Confusion Matrix and ROC-AUC Curve. The results of testing the banknote classification model show the effectiveness of the KNN algorithm model and the Gabor Filter method, as well as the assistance of PCA, producing the best performance with an accuracy value of 97.14%, precision of 95.72%, recall of 95.77%, f1-score of 95.74%, and specificity of 99.62%. The AUC value obtained on the ROC-AUC curve based on the test results produced a banknote classification model with an average AUC performance for all classes of 97.35%, which is classified as excellent in classifying banknotes, so that the model can be implemented into the system.]]>
Copyrights © 2025
