Saved in:
Bibliographic Details
Main Author: Pérez Contreras, Benjamín Felipe
Format: Recurso digital
Language:
Published: Zenodo 2026
Online Access:https://doi.org/10.5281/zenodo.18435926
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866901965235027968
author Pérez Contreras, Benjamín Felipe
author_facet Pérez Contreras, Benjamín Felipe
contents <p>e present a geometric principle demonstrating that learnability across supervised and one-class learning paradigms is governed by a single fundamental constraint: the<br>comparison of local versus global structure in the input space. Starting from first principles specifically, the Locality Axiom and the Information Axiom—we derive three domain-specific metrics: ρclass for classification, ρreg for regression, and e for one-class anomaly detection. We show theoretically and empirically that these metrics instantiate a common geometric principle:<br>                                                                                                        <strong> Learning succeeds ⇔ Local structure differs from Global structure</strong></p> <p>Extensive empirical validation on 552 real classification datasets from OpenML (+0.903 Spearman correlation), 1,452 real regression datasets (identity theorem holds in 87% of cases), and 373 anomaly detection experiments confirms strong predictive accuracy. We further demonstrate algorithm-agnosticity through validation on SVM, Random Forest, Logistic Regression, and MLP across 100 datasets. These results establish Local Coherence as a fundamental principle describing intrinsic geometric constraints on learnability.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_18435926
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle The Principle of Local Coherence A Geometric Foundation for Learnability Analysis in Classification, Regression, and Anomaly Detection
Pérez Contreras, Benjamín Felipe
<p>e present a geometric principle demonstrating that learnability across supervised and one-class learning paradigms is governed by a single fundamental constraint: the<br>comparison of local versus global structure in the input space. Starting from first principles specifically, the Locality Axiom and the Information Axiom—we derive three domain-specific metrics: ρclass for classification, ρreg for regression, and e for one-class anomaly detection. We show theoretically and empirically that these metrics instantiate a common geometric principle:<br>                                                                                                        <strong> Learning succeeds ⇔ Local structure differs from Global structure</strong></p> <p>Extensive empirical validation on 552 real classification datasets from OpenML (+0.903 Spearman correlation), 1,452 real regression datasets (identity theorem holds in 87% of cases), and 373 anomaly detection experiments confirms strong predictive accuracy. We further demonstrate algorithm-agnosticity through validation on SVM, Random Forest, Logistic Regression, and MLP across 100 datasets. These results establish Local Coherence as a fundamental principle describing intrinsic geometric constraints on learnability.</p>
title The Principle of Local Coherence A Geometric Foundation for Learnability Analysis in Classification, Regression, and Anomaly Detection
url https://doi.org/10.5281/zenodo.18435926