A project to explore Wavelet data to distinguish between real and fake banknotes.

As part of the University of London's Specialization on K-Means Clustering algorithms, I was required to build a model that could be able to distinguish between fake and real banknotes. After building the model, I was unhappy with the result. I explored two alternative techniques - Logistic Regression and Support Vector Machines (SVM). These models perform better on binary classification problems.

I created a model of each, and evaluated the results. In the end, the SVM won.

• Numpy
• Pandas
• Matplotlib
• Scipy
• Sklearn

The main file, Banknote Autentication Project.ipynb, is a Jupyter notebook. All the code and outputs are displayed for ease of use.

The data files included are the Banknote-authentication-dataset.csv and Banknote Authentication GT Labels.csv. The first is the original dataset provided by the University of London and the second is the original dataset from OpenML.