Credit-Card-Fraud-Detection

An Unsupervised Learning Approach

Experimentation notebook for the Kaggle Credit Card Fraud Detection.

Dependencies

1. matplotlib

pip3 install -U matplotlib

2. numpy

pip3 install -U numpy

3. scikit-learn

pip3 install -U scikit-learn

4. keras

pip3 install -U keras

Notebook Contents

1. Data Preprocessing

2. Data Visualization (TSNE and PCA with TSNE)

3. Create of Train, Test, Validation datasets

4. Machine Learning Approach

• K-Means Clustering
• K-Means Evaluation

6. Deep Learning Approach

• Deep AutoEncoders
• Deep AutoEncoders Evaluation

References

1. K-Means: https://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html

2. TSNE Paper: https://lvdmaaten.github.io/publications/papers/JMLR_2008.pdf

3. PCA Paper: http://pca.narod.ru/pearson1901.pdf

4. Deep AutoEncoders http://www.cs.toronto.edu/~hinton/science.pdf

Contacts

LinkedIn
Twitter
Kaggle
GitHub

License

This project is licensed under the MIT License.