credit-card-fraud-detection-using-tensor-flow

Splitting data into 4 sets

1. Shuffle/randomize data

2. One-hot encoding

3. Normalize

4. Splitting up X/y values

5. Convert data_frames to numpy arrays (float32)

6. Splitting the final data into X/y train/test

Dependencies #tensorflow #pandas #numpy

#pycharm environment is used

#Current loss: 0.8012 #Final accuracy: 99.61% #Final fraud specific accuracy: 77.67%