dirichlet_precision_wicker: Estimate Dirichlet precision using Wicker's formula

In kylebittinger/polyafit: Fit observed data to a multivariate Polya distribution

Estimate Dirichlet precision using Wicker's formula

Description

Returns an estimate of the Dirichlet precision (sum of the parameters of a Dirichlet distribution) using a closed-form expression derived by Wicker et. al.

Usage

dirichlet_precision_wicker(props, min_prop = 1e-10, min_value = 0.1)

Arguments

props	A matrix of training proportions, one column per category, one row per trial.
min_prop	Minimum proportion used in the calculation. Wicker's formula takes the log of each proportion, which fails if a proportion is zero. This parameter provides a lower bound to proportions used in the formula.
min_value	Minimum return value. In some cases, Wicker's estimate is zero or negative. This parameter is used to provide a lower bound to the return value.

Value

Estimated sum of parameters for Dirichlet distribution

References

Wicker, N., et al. "A Maximum Likelihood Approximation Method for Dirichlet's Parameter Estimation," Comp. Stat. Data Anal. 52, 1315 (2008).

kylebittinger/polyafit documentation built on Jan. 11, 2023, 8:53 a.m.