SHELF: Tools for Eliciting Probability Distributions from Experts

An R package implementing methods for eliciting probability distributions from one or more experts. It fits parametric distributions to expert-provided probability judgements and supports multivariate, Dirichlet, and meta-analysis models. The package includes R Shiny applications for interactive use and was authored by Jeremy Oakley.

Use Cases

• Fit parametric distributions based on expert-provided cumulative probability points.
• Combine judgements from multiple experts using a weighted linear pool.
• Elicit beliefs about population distributions for hierarchical models.
• Construct multivariate distributions using a Gaussian copula method.
• Elicit Dirichlet distributions for categorical data analysis.

Strengths

• Implements the established Sheffield Elicitation Framework methodology.
• Includes R Shiny apps for interactive, user-friendly application of most methods.
• Supports a range of distribution types including multivariate and for meta-analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset scale are unknown, which may limit suitability assessment.

Provenance

Source

Jeremy Oakley

Collection Method

Software package implementing statistical elicitation methods.

Time Range

null

Freshness

Last update date is unknown; freshness unverified.

Geography

null