GET: Global Envelopes for Multivariate and Functional Statistical Inference

An implementation of global envelopes, a statistical method for constructing simultaneous confidence bands for d-dimensional vectors or functions. The method, developed by Mari Myllymäki and collaborators, is described in multiple peer-reviewed publications from 2016 to 2024. It provides a framework for graphical Monte Carlo and permutation tests, central region estimation, and confidence or prediction bands.

Use Cases

• Goodness-of-fit testing for point patterns and random sets based on the described Monte Carlo test framework.
• Functional analysis of variance (ANOVA) and general linear models using the multivariate test statistic approach.
• Outlier detection and functional boxplot creation for functional data based on the central region methodology.
• Constructing global confidence and prediction bands for regression models, such as polynomial regression.
• Performing n-sample tests for correspondence of distribution functions using the permutation test capability.

Strengths

• Method is grounded in multiple peer-reviewed publications spanning from 2016 to 2024.
• Implementation supports a wide range of applications including hypothesis testing, central regions, and confidence bands.
• The core concept of a 100(1-alpha)% global envelope provides a clear probabilistic interpretation for simultaneous inference.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Mari Myllymäki

Collection Method

Implementation of a statistical method, likely code or algorithmic output.

Time Range

Method described in publications from 2016 to 2024.

Freshness

Publication references extend to 2024, but the dataset's own update date is unknown.

Geography

Global (methodological, not tied to a specific region).