The Split Is a Property of Parity, Not of Schur Complements

{"Building":[0],"Gram,":[1],"covariance,":[2],"kernel,":[3],"or":[4,9],"Toeplitz":[5],"matrices":[6],"from":[7,31,72],"exponentials":[8],"hyperbolic":[10],"functions?":[11],"You're":[12],"destroying":[13],"12":[14],"digits":[15],"before":[16],"you":[17,47,129],"factorise":[18],"anything.":[19],"One":[20],"line":[21,62],"fixes":[22],"it":[23,98,126,133],"—":[24],"34,000×":[25],"better.":[26],"When":[27],"your":[28,54,89],"entries":[29],"come":[30],"differences":[32],"of":[33,42,63,85],"large-argument":[34],"exponentials,":[35],"the":[36,45,51,57,67,79,95,108],"intermediate":[37],"products":[38],"run":[39],"13":[40],"orders":[41],"magnitude":[43],"above":[44],"values":[46],"actually":[48,118],"want,":[49],"and":[50,65,122,134],"subtraction":[52],"eats":[53],"precision.":[55],"Centring":[56],"argument":[58],"first":[59],"costs":[60],"one":[61,125],"code":[64],"cuts":[66],"worst":[68],"entry-wise":[69],"relative":[70],"error":[71],"240%":[73],"down":[74],"to":[75,131],"0.0073%.":[76],"Measured":[77],"against":[78],"largest":[80],"entry,":[81],"it's":[82],"a":[83],"factor":[84],"300":[86],"million.":[87],"If":[88],"matrix":[90],"is":[91,103,117,127],"\\"mysteriously\\"":[92],"indefinite":[93],"when":[94,130,135],"math":[96],"says":[97],"must":[99],"be":[100],"positive,":[101],"this":[102],"very":[104],"likely":[105],"why.":[106],"Separately:":[107],"block":[109],"structure":[110],"that":[111],"gets":[112],"blamed":[113],"on":[114],"Schur":[115],"complements":[116],"caused":[119],"by":[120],"parity,":[121],"knowing":[123],"which":[124],"tells":[128],"expect":[132],"not":[136],"to.":[137]}

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-09-15
DOI
https://doi.org/10.5281/zenodo.22783256
Primary Topic
Probability and Statistical Research
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
preprint

The Split Is a Property of Parity, Not of Schur Complements

Ray David Crowell
Zenodo (CERN European Organization for Nuclear Research)
Probability and Statistical Research
preprint

The Split Is a Property of Parity, Not of Schur Complements

Ray David Crowell
preprint en

Abstract

Building Gram, covariance, kernel, or Toeplitz matrices from exponentials or hyperbolic functions? You're destroying 12 digits before you factorise anything. One line fixes it — 34,000× better. When your entries come from differences of large-argument exponentials, the intermediate products run 13 orders of magnitude above the values you actually want, and the subtraction eats your precision. Centring the argument first costs one line of code and cuts the worst entry-wise relative error from 240% down to 0.0073%. Measured against the largest entry, it's a factor of 300 million. If your matrix is "mysteriously" indefinite when the math says it must be positive, this is very likely why. Separately: the block structure that gets blamed on Schur complements is actually caused by parity, and knowing which one it is tells you when to expect it and when not to.

Zenodo (CERN European Organization for Nuclear Research)
Cognizant (United States) (US)
Probability and Statistical Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.