Package: SOMevidence 1.2.0

Shaowen Ye

SOMevidence: Evidence Workflows for Self-Organizing Map Classification

Provides explicit data-design objects, leakage-safe preprocessing, design-aware resampling, reproducible ensembles of self-organizing maps, and evidence-oriented diagnostics for ecological, environmental and evolutionary applications. The package separates continuous-representation diagnostics from the defensibility of a hard partition and supports explicit abstention when evidence is insufficient. Self-organizing maps follow Kohonen (1982) <doi:10.1007/BF00337288>, using the 'kohonen' implementation described by Wehrens and Kruisselbrink (2018) <doi:10.18637/jss.v087.i07>.

Authors:Shaowen Ye [aut, cre]

SOMevidence_1.2.0.tar.gz
SOMevidence_1.2.0.zip(r-4.7-any)SOMevidence_1.2.0.zip(r-4.6-any)SOMevidence_1.2.0.zip(r-4.5-any)
SOMevidence_1.2.0.tgz(r-4.6-any)SOMevidence_1.2.0.tgz(r-4.5-any)
SOMevidence_1.2.0.tar.gz(r-4.7-any)SOMevidence_1.2.0.tar.gz(r-4.6-any)
SOMevidence_1.2.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
SOMevidence/json (API)

# Install 'SOMevidence' in R:
install.packages('SOMevidence', repos = c('https://shaowen-ye.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/shaowen-ye/somevidence/issues

Pkgdown/docs site:https://shaowen-ye.github.io

On CRAN:

Conda:

clusteringecologyenvironmental-monitoringreproducible-researchself-organizing-map

3.40 score 1 stars 25 exports 5 dependencies

Last updated from:dfc9330d33 (on v1.2.0). Checks:9 OK. Indexed: yes.

TargetResultTimeFilesSyslog
linux-develOK146
source / vignettesOK183
linux-releaseOK168
macos-releaseOK206
macos-oldrelOK229
windows-develOK120
windows-releaseOK119
windows-oldrelOK108
wasm-releaseOK100

Exports:assess_defensibilityaudit_somaudit_som_representationaudit_transfercompare_cross_modelsconsensus_somevaluate_external_labelsexpand_som_specfit_cross_modelsfit_som_ensemblelaunch_som_appmap_som_ensembleopen_data_registrypareto_candidatespartition_somplot_som_planerun_som_sensitivityrun_som_workflowsimulate_som_scenariosom_datasom_gatesom_preprocesssom_resamplessom_specupgrade_som_object

Dependencies:clueclusterkohonenRcppwithr

A real-data audit of Palmer penguin morphology
Purpose and inferential boundary | Build a design-explicit data object | Prespecify the transfer design and ensemble budget | Fit once and retain separate evidence streams | Inspect external labels only after consensus | Reproducibility record
Rendered frompalmer-penguins.Rmdusingknitr::rmarkdown

Last update: 2026-08-11
Started: 2026-08-11

API lifecycle and reproducibility contracts
Why the package has contracts | What changed in the 1.2.0 contract | What changed in the 1.1.0 contract | Stable and experimental interfaces | Object boundaries are scientific boundaries | Reporting an analysis
Rendered fromapi-lifecycle.Rmdusingknitr::rmarkdown

Last update: 2026-08-11
Started: 2026-08-10

Auditing a SOM ensemble without forcing classes
The question comes before the class map | Preserve the sampling design | Audit representation and partitions separately | A conclusion may be withheld | Governed real-data examples | Compare prespecified analysis choices on shared samples
Rendered fromgetting-started.Rmdusingknitr::rmarkdown

Last update: 2026-08-11
Started: 2026-08-10

Visualizing SOM structure and uncertainty
A map is one member of an ensemble | Ensemble evidence belongs beside the map
Rendered fromvisualizing-evidence.Rmdusingknitr::rmarkdown

Last update: 2026-08-11
Started: 2026-08-10

Respecting groups and monitoring domains in SOM validation
Why the resampling unit matters | Mapping to a held-out monitoring domain | Map a later monitoring batch without refitting | Predefined temporal or spatial blocks
Rendered fromdesign-aware-resampling.Rmdusingknitr::rmarkdown

Last update: 2026-08-10
Started: 2026-08-10

Readme and manuals

Help Manual

Help pageTopics
Assess partition defensibility without forcing a verdictassess_defensibility
Audit representation quality across a SOM ensembleaudit_som
Audit reproducibility of continuous SOM representationsaudit_som_representation
Audit transfer to held-out sampling domainsaudit_transfer
Compare SOM partitions with controlled cross-model referencescompare_cross_models
Construct a sample-level consensus partitionconsensus_som
Compare a consensus partition with external ecological labelsevaluate_external_labels
Expand an ensemble specification into its model budgetexpand_som_spec
Fit controlled cross-model reference partitionsfit_cross_models
Fit a reproducible ensemble of self-organizing mapsfit_som_ensemble
Launch the optional reproducible SOM interfacelaunch_som_app
Map new observations through a fitted SOM ensemblemap_som_ensemble
List governed tutorial and candidate open-data sourcesopen_data_registry
Identify Pareto-efficient SOM candidatespareto_candidates
Form and compare candidate hard partitions of SOM unitspartition_som
Plot an interpretable plane from one SOM ensemble memberplot_som_plane
Plot SOM audit trade-offsplot.som_audit
Plot sample-level consensus evidenceplot.som_consensus
Plot cross-model agreement without ranking methodsplot.som_cross_comparison
Plot new-data mapping shift across an ensembleplot.som_newdata_mapping
Plot partition reproducibility across candidate class countsplot.som_partitions
Plot sample-level sensitivity across prespecified scenariosplot.som_sensitivity
Plot mapping shift in held-out domainsplot.som_transfer_audit
Run prespecified SOM sensitivity scenariosrun_som_sensitivity
Run a complete SOM defensibility workflowrun_som_workflow
Simulate known-structure data for SOM validationsimulate_som_scenario
Construct a design-explicit SOM data objectsom_data
Define explicit evidence gates for a hard SOM partitionsom_gate
Define leakage-safe SOM preprocessingsom_preprocess
Construct design-aware resamplessom_resamples
Specify a reproducible SOM ensemblesom_spec
Summarize workflow quality assurance and model outcomessummary.som_workflow
Upgrade a persisted SOMevidence object contractupgrade_som_object