Credit-Card-Fraud-Detection

Using Logistic Regression and classifiers of the dataset to produce an accurate recall, f-1 and precision score Different Kernels are used to detect whether a transaction is normal or fraud. The goals of this project are:

1. Understand the distribution of the data provided
2. create a 50/50 split dataset of Fraud and Non-Fraud transactions
3. Determine the classifier to be used and finding out which has the highest accuracy
4. Create a Neural network and compare the efficiency to our best classifier
5. Understand common mistakes made with an imbalanced dataset. Imbalanced dataset requires f-1 score, confusion matrix or precision/recall score