TractaML _{- taking subjectivity out of ML}

This package aims at solving some of the most important problems that machine learning (ML) practitioners face on a daily basis, but still use sub-optimal methods to tackle them. Having this broad vision in mind, the first version of this package solves one such problem – Combined hyper-parameter tuning and feature selection.

Installation

To install the package using pip:

pip install tracta-ml

Documentation

Please refer to documentation.md for more details on package functionality.

Sample code

The example.py code presents a working example on the Wisconsin Breast Cancer (Diagnostic) dataset.

Authors

• Sanketh Nagarajan (sanketh.objml@gmail.com)

License

This project is licensed under the 3-Clause BSD License - see the LICENSE.md file for details

To cite

Please use the following format to cite this package:

N. Sanketh, "TractaML - taking subjectivity out of ML", (2018), GitHub, https://github.com/SankethNagarajan/tracta_ml

Keywords

• Combined hyper-parameter tuning and feature selection
• Micro GA
• Optimizing machine learning models
• Automatic model tuning