Ece Saltürk

Predicting fraudulent credit card transactions

Started Nov. 28, 2022

Abstract or project description

Ece has identified a public dataset that contains information about anonymized credit card transactions and whether they were fraudulent or not. The goal of this project is to build a machine learning model that can accurately predict whether or not a credit card transaction is fraudulent based on data like that found in this dataset. She is learning many of the basics of machine learning which include importing and cleaning data, correcting class imbalance and skewed variables, implementing cross-validation, and iterating through different models to find an optimal balance in the bias vs. variance tradeoff. She hopes to be able to take her findings and build it into a simplified web app using StreamLit within Python..