Skip to content

Evaluation and visualization tools for pairwise measures. The motivation for this repo is to evaluate clustering algorithms. However, amongst other use cases, this code-base is relevant in problems that involve sample pairs and distance matrices.

License

Notifications You must be signed in to change notification settings

visionjo/pairwise_metrics

Repository files navigation

Pairwise Metrics and Visualization

ql-1

Table of Contents

  • [Pairwise Metrics and Visualization](#Pairwise Metrics and Visualization)

Overview

Motivation

Use Cases

Features

Project Motivation

Additional Info

Overview

Evaluation and visualization tools for pairwise measures. The motivation for this repo is to evaluate clustering algorithms. However, amongst other use cases, this code-base is relevant in problems that involve sample pairs and distance matrices.

Motivation

SKLearn is complete with many metrics, which include submodules for clustering and pairwise. However, this code-base sets out to compliment this with pairwise metrics determined for cluster (or class) assignments of arbitrary assignments.

Use Cases

Case 1: Cluster Evaluation

Measure performance for cluster assignments (i.e., pseudo-labels) provided ground-truth labels.

Various pair-wise metrics allow for in depth analysis of clustering algorithms.

Case 2: Visualize Pairwise Relationships

Powerful visualizations generated on-the-fly provide quick and systematic way to analyze and communicate results.

Visualizations tools and demos included.

Screenshots

TODO

Tech/framework used

Ex. -

Built with

Features

TODO

Code Example

TODO

Installation

TODO

Tests

TODO

How to use?

TODO

Contribute

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.

Related Links

Action Items

This to do. Unless specified, order is arbitrary.

  • Update LICENSE
  • Complete README
  • Utility functions
  • Visualization Tools
  • Make metrics a class
    • inherent visualization tools
  • Demos and notebooks
  • Tests
    • Choose interface (i.e., python package) and create skeleton code-base
    • Unit tests
    • Type checking
      • Inputs
      • Outputs
    • All tests
    • Add script for auto-testing
      • Decide on interface/ technology for this
      • Create material providing/ setting relevant configurations
  • Add pre-commit for git (i.e., run tests before commit, and only do if tests PASS)
  • Contributing Guidelines (create)

Resources

Changelog

See changelog.md

License

MIT © Joseph Robinson

About

Evaluation and visualization tools for pairwise measures. The motivation for this repo is to evaluate clustering algorithms. However, amongst other use cases, this code-base is relevant in problems that involve sample pairs and distance matrices.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published