Welcome to SHARC’s documentation!

SHARC (SHArpened Dimensionality Reduction & Classification) is a Python library for performing local gradient clustering (LGC) based sharpened dimensionality reduction (SDR) using neural network projections (NNP) and constructing classifiers that use these projections to make classifications. The library also contains functions for finding the optimal SDR parameters and for consolidating classification results obtained through multiple classifiers.

Installation

Before installing SHARC make sure pySDR is installed first. To install pySDR please follow its respective installation instructions.

To install SHARC go inside the project directory (i.e. the directory containing the setup.py file) and run:

$ pip install .

Contents

• API
  • SDR Optimization Routines
  • SDR Optimization Metrics
  • NNP Models
  • NNP Training Utilities
  • Loss Function Definitions
  • SDR-NNP Classifier
  • Consolidation Methods
  • Additional Utilities
  • Plot Functions