summary.Powder.Individual: Estimates marginal likelihood from power posterior samples
In jeff324/powder: Power posterior sampling via differential evolution

Description Usage Arguments Value Examples

Estimates marginal likelihood using output from powder

1 2	## S3 method for class 'Powder.Individual' summary(object, options = list(), ...)

object

A Powder.Individual or Powder.Hierarchical object from powder

A tibble. If the Powder.* object was generated via the standard or parallel method from powder then the returned value will be a tibble with the following rows:

TI: log marginal likelihood estimate based on thermodynamic integration
TI Corrected: log marginal likelihood estimate based on modified thermodynamic integration using variance of log likelihood
Harmonic Mean: log marginal likelihood estimate based on harmonic mean
Steppingstone: marginal likelihood esimate based on steppingstone estimator. Because of underflow, this will usually not be meaningful.
Log Steppingstone: log marginal likelihood estimate based on steppingstone estimator.
TI Variance: estimated variance of TI
Steppingstone Variance: estimated variance of Log Steppingstone

If the object was generated via the posterior method from powder then the returned value will be a tibble with the following columns:

mean: the mean of the parameter
95% HDI UB: upper bound of 95% highest density interval
95% HDI LB: lower bound of 95% highest density interval

## Not run: 
model = LBA.Individual$new()
data('individual',package='powder')
# low number of samples for illustration purposes only
pow.out = powder(data=individual, model=model, method='parallel',burnin=100,n.samples=500)
summary(pow.out,options=list(sig.digits=5))

# A tibble: 7 x 2
# Method                       Value
# <fct>                        <dbl>
# 1 TI                     5.6759e+  2
# 2 TI Corrected           5.6799e+  2
# 3 Harmonic Mean          5.6184e+  2
# 4 Steppingstone          6.1592e+245
# 5 Log Steppingstone      5.6955e+  2
# 6 TI Variance            3.6981e-  2
# 7 Steppingstone Variance 9.8905e-  2

# do posterior estimation
pow.out = powder(data=individual, model=model, method='posterior',burnin=100,n.samples=500)
summary(pow.out,options=list(sig.digits=3))

#   parameter  mean   95% HDI LB   95% HDI UB
#   <fct>     <dbl>        <dbl>        <dbl>
# 1 A         0.827        0.553        1.02
# 2 b         0.575        0.261        1.01
# 3 sve       1.18         0.876        1.46
# 4 t0        0.269        0.178        0.320
# 5 vc        3.44         2.95         3.90
# 6 ve        1.12         0.655        1.60

## End(Not run)