•  
  •  
 

Abstract

In recent scenario detection of fakenote has become a genuine problem in the area of the financial sector as per the of various countries. In this paper, we have proposed a machine learning model that is capable of eradicating the fake note problem. In this paper, we have used a dataset of fake note images having a size of 1500. Hence exhaustive experiments have been conducted using various machine learning algorithms for proper authentication of the banknote. Here we considered K-Nearest Neighbour, Naive Bayes and random forest classifier technique yielding various result in terms of accuracy, precision and recall and f-score. It is observed that the K- nearest neighbour technique shows better performance compared to the other applied algorithm having an accuracy of 99%. Moreover, it is observed that it gives a result on determining whether a note is fake or real by output 0 when the note is fake and it gives output 1 when the note is real. Hence Knearest neighbour gives there result more accurately than other classifiers. The rules are given by machine learning classifier techniques also tested and found that they are accurate enough to be used for prediction and compare their performance to see which classifier performs best on determining the fakenote and showing their performance by bar-graphrepresentation.

Download

Share

COinS