Build a baseline whose preprocessing can be audited
Does a more complicated classifier improve on a simple baseline under the same frozen split and training-only normalization?
Make this plan your own ↓Read, edit and export without an account. This is preparation; no run or result is claimed.
Make the question your own.
6 sources · Exact preparation context
- Iris →
Attribute the supplied creators and UCI dataset DOI/source. Preserve license links and identify changes before any permitted data reuse.
Record revision: a383b6238f79de3f14ca140f4563436a9e0e31edd501edfae7c1b51e5e06f7a2
- Wine →
Attribute the supplied creators and UCI dataset DOI/source. Preserve license links and identify changes before any permitted data reuse.
Record revision: 3ece7a0b6bfc5e4ad3a2a43981123b872529b0c4625d8a291acb10ebeb760cfb
- Seeds →
Attribute the supplied creators and UCI dataset DOI/source. Preserve license links and identify changes before any permitted data reuse.
Record revision: b4c8c9782360da343c5aafef2e77b17628a5df375f3497a024a96f2928954ebf
- 3.1. Cross-validation: evaluating estimator performance →
scikit-learn documentation. 3.1. Cross-validation: evaluating estimator performance. https://scikit-learn.org/stable/modules/cross_validation.html; observed version 1.9.1 (observed documentation version; stable URL is mutable).
Record revision: 8c31ebc272585f1e00b12cc000e755dfa687ab3dba882effad7dd1a7d5ea149a
- 3.4. Metrics and scoring: quantifying the quality of predictions →
scikit-learn documentation. 3.4. Metrics and scoring: quantifying the quality of predictions. https://scikit-learn.org/stable/modules/model_evaluation.html; observed version 1.9.1 (observed documentation version; stable URL is mutable).
Record revision: 14c631ef986b838e43bdf04a5e3fd3a3be03cb58771e82dd8b05cc39b4aa6282
- scikit-learn →
Preserve the upstream project name, version, repository link, applicable notices, and contributor attribution when preparing an artifact for reuse.
Record revision: e3c10569441ff058c4f93b3456ddff6f8fc94b4bf7b065608d674a944ac15f60
Source links open the current listing. This plan retains the revisions above; listing review does not grant permission to copy or adapt source files.
Before carrying out this plan
- Payload inspection, schema validation and byte pinning remain to be done remotely.
- Package-backed experiments require qualified dependencies; no classifier has run here.
Limits to keep in view
- These small datasets do not establish state-of-the-art model performance.
- License statements are scoped observations, not blanket redistribution clearance.
Edit freely. Export a copy before reloading or leaving. Your text stays in this tab until you choose to copy, download or save it privately.
What to compare
Use identical held-out observations for paired comparisons. Select the Iris file variant explicitly, inspect Seeds feature-count metadata, and never fit normalization on the test set.
Useful outputs
- One named dataset representation, frozen stratified split and preprocessing fit boundary.
- Majority and nearest-centroid baselines with classwise metrics.
- A comparison table retaining all specified splits/seeds and uncertainty assumptions.
Inputs and prerequisites
- Payload inspection, schema validation and byte pinning remain to be done remotely.
- Package-backed experiments require qualified dependencies; no classifier has run here.
What this work would not establish
- These small datasets do not establish state-of-the-art model performance.
- License statements are scoped observations, not blanket redistribution clearance.
Start with these sources.
Iris
UCI Machine Learning Repository
Flower measurements for a compact, interpretable multiclass baseline and leakage audit.
Investigate How much do iris classification results change when preprocessing is restricted to training folds?
Wine
UCI Machine Learning Repository
Wine chemical measurements for an interpretable small-sample classification baseline.
Investigate Which wine measurements remain useful when feature selection occurs entirely inside training folds?
Seeds
UCI Machine Learning Repository
Wheat-kernel geometry for an interpretable agricultural classification baseline.
Investigate How stable are seed-class boundaries under feature scaling and small-sample resampling?
3.1. Cross-validation: evaluating estimator performance
scikit-learn documentation
A guide to model-evaluation splits, including grouped and time-series data, nested model selection and permutation-based assessment.
Investigate How much optimism arises when repeated observations from the same group are split across train and test?
3.4. Metrics and scoring: quantifying the quality of predictions
scikit-learn documentation
A broad scoring reference that distinguishes classification, regression, ranking and clustering metrics and their scorer interfaces.
Investigate Can a scoring implementation reproduce independently computed confusion-matrix and probability-loss examples, including undefined edge cases?
scikit-learn
Upstream project repository
Machine-learning estimators, preprocessing tools, model-selection routines, and evaluation metrics.
Investigate How much does fitting preprocessing before cross-validation bias a fixed synthetic prediction experiment?