Name		Name	Last commit message	Last commit date
Latest commit History 15 Commits
FCNN_SVM.py		FCNN_SVM.py
FCNN_softmax.py		FCNN_softmax.py
NN_perform.py		NN_perform.py
README.md		README.md
random_forest.py		random_forest.py
read_data.py		read_data.py
sample.py		sample.py

Repository files navigation

STAD68 Project

Dependencies

python 3
tensorflow
numpy
matplotlib
scikit-learn

Preparing data

download data from https://www.kaggle.com/dalpozz/creditcardfraud and extract
run read_data.py to store data in numpy arrays

Training

Neural Net with Cross-entropy

run FCNN_softmax.py

Neural Net with SVM layer

run FCNN_SVM.py

Random Forest

run random_forest.py. Note this script does both training and testing

Performance on test set

Please run random_forest.py first, it calculates precision-recall scores that will be used in NN_perform.py to plot precision recall curves. It also prints aucpr for the random forest along with confidence intervals.
For the neural networks run NN_perform.py, it prints aucpr for the two models, and plot precision recall curve for all models.

About

No description, website, or topics provided.

Report repository

Releases

No releases published

Packages

Contributors

Languages

Python 100.0%