Bachelor's final project.
This project is optimised for python 3+
This Data Science-like project is aimed on analyzing a bunch of techniques for False Negatives (Frauds classified as Non-Frauds) in a extreme imbalanced dataset - Financial Fraud and improving the predictors performance.
For developers, python requirements could be find in the project's root. For installing the requirements, in your venv or anaconda env, just run the following command:
pip install -r requirements.txt
.
└── mooncake
├── data
│ ├── entropy
│ │ ├── tree_entropy_2020-03-04.png
│ │ └── tree_entropy_2020-03-05.png
│ ├── datasource.csv
│ ├── validation-2020-02-27.csv
│ └── validation-2020-03-24.csv
├── docs
│ ├── confusion-matrix.xlsx
│ ├── Ficha.pdf
│ ├── PGT 04.17.docx
│ ├── PGT 05.11.docx
│ ├── PGT 05.23.docx
│ ├── PGT 05.26.docx
│ ├── PGT 05.28.docx
│ ├── PGT 05.29.docx
│ └── PGT 05.29.pdf
├── mooncake
│ ├── __init__.py
│ ├── helpers.py
│ ├── models.py
│ ├── plotting.ipynb
│ ├── ros.py
│ ├── smote.ipynb
│ └── smote.py
├── tests
│ └── unittests
│ ├── __init__.py
│ ├── ...
│ └── test_helpers.py
├── .gitignore
├── LICENSE
├── README.md
├── requirements.txt
└── setup.py
-
Activation Functions: link;
-
ANN in java: Sorting_Hat
-
Benchmark's article: Expertise areas
-
Old Project's name inspiration: Tree-Shrews;