Skip to content

Latest commit

 

History

9 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MeasureQual

Claim-scoped qualification of computational measurements

MeasureQual is an R framework for determining whether a computational measurement is qualified for a specific scientific claim under a defined design, context, and reference structure. Analysts declare assumptions and decision rules; the framework checks prerequisites and constrains interpretation.

It is not a universal signature-ranking tool, biological-truth engine, leaderboard, automatic claim-validation system or causal-inference engine.

Installation

remotes::install_github("ZichenSheng/MeasureQual")
# Optional frozen signature/resource registries; independent of MeasureQual:
remotes::install_github("ZichenSheng/MeasureQual", subdir = "companion/gbm201data", force = TRUE)

The companion command uses force = TRUE to avoid remotes repository-name cache shortcuts when installing two packages from the same repository.

Synthetic example

library(MeasureQual)
source(system.file("examples", "toy-workflow.R", package = "MeasureQual"))
toy <- toy_spec()
frozen <- freeze_audit(toy$spec)
check <- check_audit(frozen, toy$data)
run <- run_audit(check, toy$data)
record <- evidence_record(run, render = FALSE)
record

The example uses synthetic observations only. Five vignettes explain reference dependence, truth isolation, unresolved decisions, definitions and constructibility. An empty decision map yields UNRESOLVED; a blocked preflight can itself produce a record. Terminal states are unordered, not a score:

  • SUPPORTED_WITHIN_DOMAIN
  • DEFINITION_IMPLIED
  • DIRECTIONALLY_MISLEADING
  • NOT_EVALUABLE
  • INVALID
  • UNRESOLVED

Scope and provenance

MeasureQual supports explicit declarations for informativeness, resolution, context deformation, detectability, reference/null construction and recovery or qualification behavior. It evaluates the analyst's specified contract; it does not supply automatic scientific validity or causal identification. The terminal states are unordered and scoped to each claim and context.

Glioblastoma transcriptomic signatures are a motivating application, not the software's scope. The optional gbm201data companion supplies frozen resource definitions for manuscript-associated audits. Raw study databases and real patient, sample, cell, spot or AOI observations are not included.

R >= 4.1 is required. See RENAME_INTEGRITY_REPORT.md for verification and MIGRATION.md for migration details. Original software is MIT licensed; no third-party source dataset is relicensed or redistributed.

Citation and history

Cite MeasureQual 0.10.0 using the version DOI or CITATION.cff. The concept DOI identifies the continuing software lineage.

Formerly released as evaluably v0.9.0, preserved at its historical DOI.

About

Claim-scoped qualification of computational measurements in R

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages