Skip to contents

Fit a Growth Model to a Bipod Object

fit.Rd

This function fits a specified growth model to a bipod object, allowing for exponential or logistic growth fitting. If both models are considered, an automatic model selection process determines the best fit.

Usage

fit(
  x,
  growth_type = "exponential",
  infer_t0 = TRUE,
  variational = FALSE,
  factor_size = 1,
  model_selection_algo = "bayes_factor",
  chains = 4,
  iter = 5000,
  cores = 4
)

Arguments

x

A bipod object containing population count data over time.

growth_type

Character string specifying the growth model to fit. Options are:

  • "exponential" – Fits an exponential growth model.

  • "logistic" – Fits a logistic growth model.

  • "both" – Performs model selection between exponential and logistic growth. Default is "exponential".

infer_t0

Logical value indicating whether to infer the initial time of population origin (t0). If TRUE, t0 is estimated as part of the model fitting. Default is TRUE.

variational

Logical value indicating whether to use variational inference instead of Markov Chain Monte Carlo (MCMC) sampling. If TRUE, variational inference is applied; otherwise, MCMC is used. Default is FALSE.

factor_size

Numeric value used to scale population counts in the bipod object. Must be positive and no larger than the minimum count value. Default is 1.

model_selection_algo

Character string specifying the model selection algorithm when growth_type = "both". Options are:

  • "bayes_factor" – Compares models using Bayes factors.

  • "mixture_model" – Uses a mixture modeling approach to estimate probabilities. Default is "bayes_factor".

chains

Integer specifying the number of MCMC chains. Ignored if variational = TRUE. Default is 4.

iter

Integer specifying the number of MCMC iterations. Ignored if variational = TRUE. Default is 5000.

cores

Integer specifying the number of CPU cores to use for parallel processing. Default is 4.

Value

Returns the input bipod object with additional attributes:

  • fit – The fitted growth model.

  • fit_info – Metadata about the fitting process, including:

    • Sampling method (MCMC or variational inference)

    • Factor size used for scaling

    • Selected growth model

    • Model selection details (if applicable)

Examples

# Create a bipod object with your data
data = biPOD::sim_stochastic_exponential(100, 1, 0, 10, .25)
x = biPOD::init(data, "sample")
#> 
#> ── biPOD - bayesian inference for Population Dynamics ──────────────────────────
#> 
#> ℹ Using sample named: sample.
x <- fit(x, growth_type = "both", model_selection_algo = "bayes_factor")
#> ℹ Fitting with model selection.
#> 
#> ℹ Model selection finished!
#> ℹ Model with "Exponential" growth deemed better with "Very strong" evidence. (BF = 90.9536467166196)
biPOD::plot_fit(x, CI = .8)