10.46243/jst.2021.v6.i04.pp285-290 registered

The currency coins or banknotes used by any country to perform economic activities in the global market should be always genuine. However, some of the miscreants in the society provoke forged notes into the bazaar which bears a resemblance to exactly the genuine note. It is stiff with human naked eye to notify the difference in between these two because they have a lot of analogous facial appearance. Hence, there is a call for of competent validation system which informs accurately whether the note in the transaction is authentic or not. This paper proposes machine learning based methodology to classify the fake and genuine notes. Accordingly, exhaustive testing has been performed using Multi Layer Perceptron Neural Network (MLPNN), Naïve Bayes and Random Forest (RF) Algorithms on the standard data set. The testing results show that RF and MLPNN algorithms gives comparable results in terms of accuracy and other performance measures as compared to any other algorithms. The complete training and testing of the dataset is performed in WEKA software.