select_best_trait_model_from_geiger: Compare trait evolutionary model fits with AICc and Akaike's...

In deepSTRAPP: Test for Differences in Diversification Rates over Time

Compare trait evolutionary model fits with AICc and Akaike's weights

Description

Compare evolutionary models fit with geiger::fitContinuous() or geiger::fitDiscrete() using AICc and Akaike's weights. Generate a data.frame summarizing information. Identify the best model and extract its results.

Usage

select_best_trait_model_from_geiger(list_model_fits)

Arguments

list_model_fits

Named list with the results of a model fit with geiger::fitContinuous() or geiger::fitDiscrete() in each element.

Value

The function returns a list with three elements.

• ⁠$model_comparison_df⁠ Data.frame summarizing information to compare model fits. It includes the model name (⁠$model⁠), the log-likelihood (⁠$logLik⁠), the number of free-parameters (⁠$k⁠), the AIC (⁠$AICc⁠), the corrected AIC (⁠$AICc⁠), the delta to the best/lowest AICc (⁠$delta_AICc⁠), the Akaike weights (⁠$Akaike_weights⁠), and their rank based on AICc (⁠$rank⁠).

• ⁠$best_model_name⁠ Character string. Name of the best model.

• ⁠$best_model_fit⁠ List containing the output of geiger::fitContinuous() or geiger::fitDiscrete() for the model with the best fit.

Author(s)

Maël Doré

See Also

geiger::fitContinuous() geiger::fitDiscrete()

Examples

# ----- Example 1: Continuous data ----- #

# Load phylogeny and tip data
library(phytools)
data(eel.tree)
data(eel.data)

# Extract body size
eel_data <- stats::setNames(eel.data$Max_TL_cm,
                            rownames(eel.data))

 # (May take several minutes to run)
# Fit BM model
BM_fit <- geiger::fitContinuous(phy = eel.tree, dat = eel_data, model = "BM")
# Fit EB model
EB_fit <- geiger::fitContinuous(phy = eel.tree, dat = eel_data, model = "EB")
# Fit OU model
OU_fit <- geiger::fitContinuous(phy = eel.tree, dat = eel_data, model = "OU")

# Store models
list_model_fits <- list(BM = BM_fit, EB = EB_fit, OU = OU_fit)

# Compare models
model_comparison_output <- select_best_trait_model_from_geiger(list_model_fits = list_model_fits)

# Explore output
str(model_comparison_output, max.level = 2)

# Print comparison
print(model_comparison_output$models_comparison_df)

# Print best model fit
print(model_comparison_output$best_model_fit) 

# ----- Example 2: Categorical data ----- #

# Load phylogeny and tip data
library(phytools)
data(eel.tree)
data(eel.data)

# Extract feeding mode data
eel_data <- stats::setNames(eel.data$feed_mode, rownames(eel.data))
table(eel_data)

 # (May take several minutes to run)
# Fit ER model
ER_fit <- geiger::fitDiscrete(phy = eel.tree, dat = eel_data, model = "ER")
# Fit ARD model
ARD_fit <- geiger::fitDiscrete(phy = eel.tree, dat = eel_data, model = "ARD")

# Store models
list_model_fits <- list(ER = ER_fit, ARD = ARD_fit)

# Compare models
model_comparison_output <- select_best_trait_model_from_geiger(list_model_fits = list_model_fits)

# Explore output
str(model_comparison_output, max.level = 2)

# Print comparison
print(model_comparison_output$models_comparison_df)

# Print best model fit
print(model_comparison_output$best_model_fit) 

deepSTRAPP documentation built on Jan. 20, 2026, 1:06 a.m.