Data

Tools

Indexes

Applications

Experiments

Open Source Science

Recent Releases of fseval

fseval - 3.1.0

What's Changed

  • Recipe/running on aws by @dunnkers in https://github.com/dunnkers/fseval/pull/86
  • Install Node.js in the Devcontainer by @dunnkers in https://github.com/dunnkers/fseval/pull/87
  • Updated schematic according to paper by @dunnkers in https://github.com/dunnkers/fseval/pull/88
  • Add Developer instructions by @dunnkers in https://github.com/dunnkers/fseval/pull/89
  • Add big feature selector benchmark by @dunnkers in https://github.com/dunnkers/fseval/pull/90
  • Add statement of need to docs by @dunnkers in https://github.com/dunnkers/fseval/pull/92

Full Changelog: https://github.com/dunnkers/fseval/compare/v3.0.4...3.1.0

Scientific Software - Peer-reviewed - Python
Published by dunnkers over 3 years ago

fseval - 3.0.4

What's Changed

  • Fix license badge by @dunnkers in https://github.com/dunnkers/fseval/pull/80
  • Improve installation instructions by @dunnkers in https://github.com/dunnkers/fseval/pull/81
  • Add some docs explaining how to "analyze FS algorithm stability" by @dunnkers in https://github.com/dunnkers/fseval/pull/83
  • Small-fix-in-algorithm-stability-plus-upgrade-docusaurus by @dunnkers in https://github.com/dunnkers/fseval/pull/84

Full Changelog: https://github.com/dunnkers/fseval/compare/v3.0.3...v3.0.4

Scientific Software - Peer-reviewed - Python
Published by dunnkers over 3 years ago

fseval - 3.0.3

  • [x] Upgrade to hydra 1.2
  • [x] Improve docs quickstart

Scientific Software - Peer-reviewed - Python
Published by dunnkers almost 4 years ago

fseval - 3.0.2

  • [x] Bump some deps. prismjs and follow-redirects.

Scientific Software - Peer-reviewed - Python
Published by dunnkers about 4 years ago

fseval - 3.0.1

Fix issues: - [x] 🐛 Fix to_sql callback compatibility with n_jobs > 1 (#64)

Scientific Software - Peer-reviewed - Python
Published by dunnkers over 4 years ago

fseval - 3.0.0

New release! 🚀

Many parts of the API changed for the sake of better documentation. Most importantly: - [x] Usage is now not primarily through a CLI, but by defining your own Hydra app. This lets the user easier understand what's going on. - [x] Classifiers/Regressors are now defined in separate configs. This makes the entire config easier to read/reason about. - [x] Docstrings for every configuration needed. - [x] Documentation website with Docusaurus. - [x] Much higher test coverage. Badges for test coverage and supported Python versions.

Checkout the new documentation website!

Scientific Software - Peer-reviewed - Python
Published by dunnkers over 4 years ago

fseval - 2.1.2

  • [x] Update Vega config fields
  • [x] Fail pipeline when a feature importance is less than zero

Scientific Software - Peer-reviewed - Python
Published by dunnkers almost 5 years ago

fseval - 2.1.1

  • [x] Better default: resample=shuffle
  • [x] Better default: n_bootstraps=1
  • [x] Remove unused dependency, 'tqdm'

Scientific Software - Peer-reviewed - Python
Published by dunnkers almost 5 years ago

fseval - 2.1.0

Many fixes and new features. - [x] 🏷 More built-ins: add 5 OpenML datasets and add Infinite Selection (b52e0dad3b45172a9a63de76e426a22505938df4 and 008f9eda8a3f51e06fe29b77ed6f411130b367ec) - [x] 💪🏻 Add multiprocessing support (53987d079244eb2604f9b4ecd76ae28ec9080dd2, cacc902881bc92bde16d72863d2ede0d51774e5f) - [x] 🌺 Feature support and feature rankings are now also scored (9ee15a55809df235a7b152f34be8e35cf3916216) - [x] 📊 Upload raw feature ranker matrices, so they can be processed on the frontend (146bc47a23d11f3de7aad55dc795b0568d415adf) - [x] 🔨 Fix caching. Now behaves like expected. (07e631d56db3bfd666f519f2816a1db6930f8ec8, 86f4ca3b552bd50c0883fb5349bb6c49ede6e3d1) - [x] ⚙️ Upgrade to Hydra 1.1.0 (8e0d976ebb4dd70143578a4621745e68e14c9688) - [x] 📊 Feature support vectors are now validated in a separate step. Outputs another validation score. (dea00ab947c5b1a8816efd3c619d4512420d5637) - [x] 📊 Add charts as part of pipeline. This allows users to get interpretable results without any configuring on their part. (f050a87cbce8a2f614c4bcabc0ee2f5f72e6cb52)

Scientific Software - Peer-reviewed - Python
Published by dunnkers almost 5 years ago

fseval - 2.0.5

  • [x] Loosen up package deps
  • [x] Upload a summary table
  • [x] Allow disabling metric logging

Scientific Software - Peer-reviewed - Python
Published by dunnkers almost 5 years ago

fseval - 2.0.4

Minor release: - [x] Fix PyPi package - [x] Don't bundle estimator dependencies

Scientific Software - Peer-reviewed - Python
Published by dunnkers almost 5 years ago

fseval - 2.0.3

Scientific Software - Peer-reviewed - Python
Published by dunnkers about 5 years ago

fseval - 2.0.2

Minor release. - [x] Adds README information to PyPi page

Scientific Software - Peer-reviewed - Python
Published by dunnkers about 5 years ago

fseval - 2.0.1

  • [x] Bugfixes
  • [x] Fix PyPi distribution

Scientific Software - Peer-reviewed - Python
Published by dunnkers about 5 years ago

fseval - 2.0.0 🚀

Completely rewritten framework, integrating wandb, hydra and sklearn for easily configurable and flexible feature selection benchmarking.

  • [x] Beautifully visualize benchmarking results inside a wandb dashboard ✨✨
  • [x] Ability to enqueue jobs in a Redis database using Hydra's RQ Launcher.
  • [x] Both classification and regression datasets supported.
  • [x] Multi-label classification and multi-output classification- and regression all supported.
  • [x] Flexible dataset loading using adapters: adapters for OpenML and wandb artifacts are included.
  • [x] Ability to dynamically define a feature importances ground-truth on any dataset - can be used to evaluate feature rankings.
  • [x] Sophisticated evaluation metrics are used to evaluate the quality of a feature ranking.

🙌🏻

Scientific Software - Peer-reviewed - Python
Published by dunnkers about 5 years ago

fseval - 1.0.0: Initial release

  • [x] Benchmark feature selection algorithms
  • [x] Complete pipeline for benchmarking: reading datasets, performing feature selection, computing evaluation metrics, plotting results
  • [x] Written from the ground up to be parallel-compatible
  • [x] SLURM interoperability (e.g. University of Groningen's Peregrine cluster)
  • [x] CLI tool for submitting jobs

Scientific Software - Peer-reviewed - Python
Published by dunnkers over 5 years ago