HARDy¶
HARDy Handling, Arbitrage, Recognition of Data using python is a
package to assist in discovery, research and classification of data.
HARDy utilizes feature engineering and kernal methods simultaneously
to improve the accuracy of convolutional neural network. HARDy enables
the user to search over given range of hyperparameters to achieve the best
prediction score for a machine learning model. The summary of results can be
printed out using single argument custom built plotly function.
Installation¶
The easiest way to install HARDy is using conda:
conda install -c pozzorg hardy
For detailed installation instructions and other methods of installation, visit Installation Page
Dependencies¶
HARDy requires following modules:
- Python (3.7)
- Tensorflow (2.0)
- Keras Tuner (1.0.0)
Complete list of dependencies can be found in the environment file