This is my implementation for the research paper "Feature engineering strategies for credit card fraud detection" by Bahnsen et. al link. In this implementation, I have tried to implement the 3rd section from the paper i.e., "Feature engineering for fraud detection". Here, authors have first proposed features for capturing the customer spending patterns, followed by time features.
Checkout the notebook Aggregated_Features.ipynb for calculating the set of features capturing the customer spending pattern and Time_Features.ipynb for time features. Also, checkout the library PyCircular by one of the paper authors and its blogpost to understand the implementation of time features in more detail.