Banknote Authentication Using Logistic Regression and Artificial Neural Networks

Banknotes are special notes authorized by the government and carry monetary value. As a result, there are incentives for criminals to create counterfeit. The goal of this study is to create models using machine learning techniques that can accurately classify a banknote as authentic or fake. The methods used were logistic regression and artificial neural networks. An open-source data set was obtained and split into 7 sub-datasets, and multiple models were created to model the data. There was a total of 7 logistic regression models, each corresponding to one of the 7 sub-datasets. Additionally, an artificial neural networks model was used on the 7th sub-dataset. Both the neural networks model and the logistic regression model achieved accuracies greater than 99%.