Curated Comparison Evidence

Curated for DeTime 0.1.3 on 2026-07-14T15:15:44.279658+00:00.

Release state: local-unverified.

Audit date: 2026-07-14 (utc-current).

Comparator versions:

  • statsmodels: 0.14.6
  • PyEMD/EMD-signal: not installed in local audit environment
  • PyWavelets: 1.9.0
  • PySDKit: 0.4.25
  • SSALib: not installed in local audit environment
  • sktime: not installed in local audit environment

Runtime and memory boundary: feature audit only; no cross-package runtime or peak-memory ranking is claimed.

Capability matrix

package primary_scope unified_config_object unified_result_object machine_readable_catalog batch_cli profiling_path multivariate_support maturity_labeling compact_output
DeTime standalone cross-family time-series decomposition yes yes yes yes yes yes explicit yes
statsmodels classical decomposition and modeling per-method constructors shared for STL/MSTL library metadata library API external tooling limited implicit method result
PyEMD EMD-family toolkit family classes IMF arrays method state library API external tooling family-specific family-specific IMF arrays
PyWavelets wavelet toolkit function arguments coefficient trees transform metadata library API external tooling transform-specific family-specific coefficients
PySDKit signal decomposition toolkit method-family classes method-specific method documentation library API external tooling yes less explicit method-specific
SSALib SSA specialist toolkit SSA-specific SSA-specific SSA metadata library API external tooling no focused SSA results
sktime broad time-series ecosystem estimator parameters task-specific runtime estimator registry library API external tooling partial ecosystem-level task-specific

Install and reproducibility matrix

package public_release_story pypi_path github_release wheels ci_platforms docs_website tutorials api_docs coverage_disclosure reproducibility_script
DeTime 0.1.3 local source state; publication status unverified not verified not verified yes ubuntu, macOS, windows GitHub Pages yes yes dual report yes
statsmodels yes yes yes yes yes yes yes yes project-dependent project-dependent
PyEMD yes yes yes partial project-dependent partial partial partial project-dependent no
PyWavelets yes yes yes yes yes yes yes yes project-dependent no
PySDKit yes yes yes partial project-dependent yes partial partial project-dependent no
SSALib yes yes yes partial project-dependent yes partial partial project-dependent no
sktime yes yes yes yes yes yes yes yes project-dependent project-dependent

Family-fairness notes

family specialist specialist_deeper_strength de_time_position
classical decomposition statsmodels deeper classical decomposition and statistical modeling wraps STL and MSTL under one config/result/CLI layer
SSA SSALib deeper SSA-only environment with SSA-specific tooling uses SSA as one core method inside a cross-family workflow layer
EMD family PyEMD deeper EMD-family tooling exposes EMD and CEEMDAN through the same contract used for other families
wavelet workflows PyWavelets deeper wavelet transforms and transform-specific APIs uses wavelets as one workflow option rather than a wavelet-first toolkit
unified toolkit layer PySDKit broader signal-decomposition toolkit with optional multivariate backends focuses on time-series decomposition workflows, machine-facing contracts, and compact outputs

Machine-contract matrix

package json_schema_assets compact_result_modes metadata_shortlist_interface machine_readable_catalog cli_schema_command artifact_contract
DeTime yes full/summary/meta yes yes yes yes
statsmodels no no no no no no
PyEMD no no no no no no
PyWavelets no no no no no no
PySDKit no no no no no no
SSALib no no no no no no
sktime no no no yes-runtime-registry no no

Runtime snapshot source

  • docs\assets\generated\evidence\performance_snapshot.csv