Data

Tools

Indexes

Applications

Experiments

Open Source Science

nescience

Machine Learning with the Minimum Nescience Principle.

https://github.com/rleiva/nescience

Science Score: 44.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ✓
    CITATION.cff file
    Found CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Committers with academic emails
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (6.4%) to scientific vocabulary
Last synced: 7 months ago · JSON representation ·

Repository

Machine Learning with the Minimum Nescience Principle.

Basic Info
  • Host: GitHub
  • Owner: rleiva
  • License: other
  • Language: Jupyter Notebook
  • Default Branch: master
  • Homepage:
  • Size: 2.84 MB
Statistics
  • Stars: 11
  • Watchers: 2
  • Forks: 6
  • Open Issues: 18
  • Releases: 1
Created over 6 years ago · Last pushed 7 months ago
Metadata Files
Readme Contributing License Citation

README.md

nescience

Machine learning guided by the Minimum Nescience Principle — pick models and features by minimizing a three-part objective: - μ (miscoding) — representation inadequacy (low MI between X and y). - ι (inaccuracy) — predictive error on fresh data. - σ (surfeit) — model superfluity/complexity.

The library provides metrics, a unified Nescience objective, and search wrappers that choose both features and estimators with scikit-learn compatibility.

Research background: R. A. Garcia Leiva, A Mathematical Theory of the Unknown (Theory of Nescience).

Quick start

bash pip install -U nescience

```python from nescience.estimators import NescienceClassifier from sklearn.datasets import loadbreastcancer X, y = loadbreastcancer(returnXy=True)

clf = NescienceClassifier(searchbudget=60, randomstate=0) clf.fit(X, y) print(clf.bestestimator) print(clf.nesciencebreakdown) ```

Philosophy (short)

  • μ: If the representation barely informs the target, you are doomed. We approximate this with normalized mutual information between features and target.
  • ι: If predictions fail on unseen data, the description is inaccurate. We estimate with task-appropriate loss.
  • σ: If the model is overly complex relative to the data, you’re probably memorizing. We approximate with a normalized description-length proxy.

Each component is pluggable.

Status

This is a reboot (v0.2.0). API may evolve before v1.0. See the roadmap in CONTRIBUTING.md.

Citation (CITATION.cff) 

cff-version: 1.2.0
message: "If you use this software, please cite it as below."
title: "nescience: Machine learning guided by the Minimum Nescience Principle"
authors:
  - family-names: "Garcia Leiva"
    given-names: "Rafael A."
date-released: "2025-08-12"
version: "0.2.0"
repository-code: "https://github.com/rleiva/nescience"
preferred-citation:
  type: book
  title: "A Mathematical Theory of the Unknown: Journey Beyond the Frontiers of Human Understanding"
  authors:
    - family-names: "Garcia Leiva"
      given-names: "Rafael A."
  year: 2025
  url: "https://leanpub.com/nescience"

GitHub Events

Total
  • Watch event: 1
  • Push event: 4
  • Pull request event: 3
  • Create event: 1
Last Year
  • Watch event: 1
  • Push event: 4
  • Pull request event: 3
  • Create event: 1

Committers

Last synced: over 2 years ago

All Time
  • Total Commits: 104
  • Total Committers: 5
  • Avg Commits per committer: 20.8
  • Development Distribution Score (DDS): 0.221
Past Year
  • Commits: 0
  • Committers: 0
  • Avg Commits per committer: 0.0
  • Development Distribution Score (DDS): 0.0
Top Committers
Name Email Commits
R. Garcia Leiva r****a@g****m 81
Rafael Garcia Leiva r****a@i****g 14
simon814b 4****b 5
simon814b s****u@g****m 3
mlt m****0@g****m 1
Committer Domains (Top 20 + Academic)
imdea.org: 1

Issues and Pull Requests

Last synced: 7 months ago

All Time
  • Total issues: 25
  • Total pull requests: 5
  • Average time to close issues: 4 months
  • Average time to close pull requests: 4 days
  • Total issue authors: 2
  • Total pull request authors: 3
  • Average comments per issue: 0.24
  • Average comments per pull request: 0.0
  • Merged pull requests: 4
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 2
  • Average time to close issues: N/A
  • Average time to close pull requests: less than a minute
  • Issue authors: 0
  • Pull request authors: 1
  • Average comments per issue: 0
  • Average comments per pull request: 0.0
  • Merged pull requests: 1
  • Bot issues: 0
  • Bot pull requests: 0
View more stats
Top Authors
Issue Authors
  • rleiva (21)
  • simon814b (4)
Pull Request Authors
  • simon814b (2)
  • rleiva (2)
  • Mohmoulay (1)
Top Labels
Issue Labels
bug (10) enhancement (3) good first issue (2) research (2)
Pull Request Labels

Packages

  • Total packages: 1
  • Total downloads:
    • pypi 5 last-month
  • Total dependent packages: 0
  • Total dependent repositories: 1
  • Total versions: 1
  • Total maintainers: 1
pypi.org: nescience

Machine Learning with the Minimum Nescience Principle

  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 1
  • Downloads: 5 Last month
Rankings
Dependent packages count: 10.0%
Forks count: 13.3%
Stargazers count: 16.5%
Dependent repos count: 21.7%
Average: 27.1%
Downloads: 73.8%
Maintainers (1)
rleiva
Last synced: 7 months ago

Dependencies

.github/workflows/ci.yml actions
  • actions/checkout v4 composite
  • actions/setup-python v5 composite
pyproject.toml pypi
  • joblib >=1.2
  • numpy >=1.23
  • scikit-learn >=1.6,<1.8
  • scipy >=1.9