Identify The Authenticity of Rupiah Currency Using K Nearest Neighbor (K-NN) Algorithm

Abstract

The rupiah currency is a valid exchange rate used in transactions in the Republic of Indonesia. The Rupiah is often falsified as paper currency. Rupiah paper has a unique texture characteristic so that if processed digitally, it will be easy to distinguish from fake ones.  Designing the authenticity of Rupiah currency system using the K-NN method aims to facilitate the authenticity of the currency and test the accuracy of the method used. The method used in this research is the method of Gray Level Co-occurrence Matrix (GLCM) as a method of feature extraction and K-Nearest Neighbor (K-NN) algorithm used in the identification process. The testing phase uses data for 18 currency images. The results showed an accuracy rate of 100% for the value k = 1, 77.78% for the value k = 3, and 55.56% for the value k = 5. The highest level of accuracy in a currency authenticity identification system occurs when the value of k = 1 is 100%. The value of k on the classification input using the K-NN can determine the level of accuracy of the classification process.