modelsupport() computes measures of the relative support of each of the best models identified by modelselect() to facilitate selecting the optimal number and position of regions. These measures are in the form of information criteria (AICc and BIC).
Arguments
- models
a
regions_modelselectobject; the output of a call tomodelselect().
Value
A regions_modelsupport object, which contains the best model for each number of regions as determined by the AICc and BIC. The computed statistics are AICc/BIC–the value of the information criterion (IC) for each model, deltaAIC/deltaBIC–the difference between the IC for the corresponding model and that of the model with the lowest IC value, model_lik–the likelihood ratio of the model against the model with the lowest IC value, and Ak_weight/BIC_weight–the Akaike weights for each model used to compute the region score. The region score is a weighted average of the numbers of regions, weighted by the Akaike weights to represent the variability around the optimal number of regions.
Examples
data("alligator")
alligator_data <- process_measurements(alligator,
pos = "Vertebra")
# Compute PCOs
alligator_PCO <- svdPCO(alligator_data)
# Fit segmented regression models for 1 to 7 regions
# using PCOs 1 to 4 and a continuous model with a
# non-exhaustive search
regionresults <- calcregions(alligator_PCO,
scores = 1:4,
noregions = 7,
minvert = 3,
cont = TRUE,
exhaus = FALSE,
verbose = FALSE)
regionresults
#> A `regions_results` object
#> - number of PCOs used: 4
#> - number of regions: 1, 2, 3, 4, 5, 6, 7
#> - model type: continuous
#> - min vertebrae per region: 3
#> - total models saved: 112
#> Use `summary()` to examine summaries of the fitting process.
# For each number of regions, identify best
# model based on minimizing RSS
bestresults <- modelselect(regionresults)
bestresults
#> Regions BP 1 BP 2 BP 3 BP 4 BP 5 BP 6 sumRSS RSS.1 RSS.2 RSS.3 RSS.4
#> 1 . . . . . . 0.725 0.221 0.327 0.135 0.041
#> 2 12 . . . . . 0.356 0.077 0.127 0.112 0.040
#> 3 9 14 . . . . 0.154 0.019 0.039 0.063 0.033
#> 4 9 13 19 . . . 0.098 0.008 0.010 0.045 0.035
#> 5 6 9 13 19 . . 0.054 0.008 0.010 0.017 0.020
#> 6 6 9 12 15 19 . 0.042 0.007 0.009 0.017 0.009
#> 7 6 9 12 15 18 21 0.042 0.006 0.011 0.014 0.010
# Evaluate support for each model and rank models
supp <- modelsupport(bestresults)
supp
#> - Model support (AICc)
#> Regions BP 1 BP 2 BP 3 BP 4 BP 5 BP 6 sumRSS AICc deltaAIC model_lik
#> 5 6 9 13 19 . . 0.054 -566.603 0.000 1.000
#> 6 6 9 12 15 19 . 0.042 -565.883 0.720 0.698
#> 7 6 9 12 15 18 21 0.042 -537.487 29.116 0.000
#> 4 9 13 19 . . . 0.098 -534.934 31.668 0.000
#> 3 9 14 . . . . 0.154 -512.839 53.764 0.000
#> 2 12 . . . . . 0.356 -454.040 112.563 0.000
#> 1 . . . . . . 0.725 -404.530 162.073 0.000
#> Ak_weight
#> 0.589
#> 0.411
#> 0.000
#> 0.000
#> 0.000
#> 0.000
#> 0.000
#> Region score: 5.41
#>
#> - Model support (BIC)
#> Regions BP 1 BP 2 BP 3 BP 4 BP 5 BP 6 sumRSS BIC deltaBIC model_lik
#> 6 6 9 12 15 19 . 0.042 -525.687 0.000 1.00
#> 5 6 9 13 19 . . 0.054 -524.763 0.924 0.63
#> 7 6 9 12 15 18 21 0.042 -503.838 21.849 0.00
#> 4 9 13 19 . . . 0.098 -495.206 30.481 0.00
#> 3 9 14 . . . . 0.154 -478.160 47.527 0.00
#> 2 12 . . . . . 0.356 -426.753 98.933 0.00
#> 1 . . . . . . 0.725 -386.534 139.152 0.00
#> BIC_weight
#> 0.613
#> 0.387
#> 0.000
#> 0.000
#> 0.000
#> 0.000
#> 0.000
#> Region score: 5.61
# 5 regions best based on AICc; 6 regions based on BIC