gammaLogMLE: GammaModel::gammaLogMLE

In UCL/GammaModel: Modelling age-at-death profiles using the Gamma distribution

Description

Finds the Maximum Likelihood Estimate for a Gamma distribution, given some age-class count data

Usage

gammaLogMLE(counts, class.ages = NULL, trace = FALSE)

Arguments

counts	A one-row data frame comprising integer counts, with column names in capital letters. Formatting requirements detailed in checkCountsFormat.
class.ages	An optional one-row data frame specifying the starting age of each age-class. Age-classes as column names.
trace	An optional logical value indicating if the search progress should be printed.

Details

class.ages must be a one-row data frame giving the starting age of each age class or age multi-class. These must be sequential, such that the start of the (n+1)th class equals the end of the nth class. Since all possible ages must be accounted for, the first starting age must be zero, whilst the final age class will automatically include ages above this final age.

If the default class.ages = NULL is used, Payne's caprine age classes are used as follows, where units are in years:

class.ages <- data.frame(A=0, B=1/6, C=1/2, D=1, E=2, F=3, G=4, H=6, I=8)

The search for the Maximum Likelihood parameters is acheived using a bespoke jDE variant of the Differential Evolution stochastic algorithm for global optimization, as implemented by JDEoptim in Eduardo L T Conceicao's DEoptimR package.

Value

Single value numeric giving the exact log likelihood')

References

Kill-off Patterns in Sheep and Goats: The Mandibles from Asvan Kale. Payne 1973.

See Also

JDEoptim

Examples

    # Maximum Likelihood Estimate of Gamma distribution
    data(Neolithic)
    counts <- Neolithic['TRA1',]
    gamma.ll <- gammaLogMLE(counts)
	

UCL/GammaModel documentation built on June 25, 2020, 5:38 a.m.