Skip to content

Latest commit

 

History

History
31 lines (15 loc) · 672 Bytes

README.md

File metadata and controls

31 lines (15 loc) · 672 Bytes
Raw

CreditFraud-Classifier

This project implements Artificial Neural Network ANN to classify valid and fake credit transaction. Firstly a 4 layer ANN model is created and trained using the ibalanced credit fraud dataset. And then the same model is trained with a balanced dataset created using SMOTETomek. Lastly a hyperparameter tuned ANN using kerastuner is created, and the model accuracies are compared

Techniques used

  1. Data Cleaning

  2. Data visualization

  3. ANN Modelling

Algorithms used

  1. Artificial Feedforward Neural Network (ANN)

Libraries/Toold required

  1. Pandas

  2. seaborn

  3. scikit learn

  4. keras

  5. imblearn

  6. tensorflow