Package index • FINN

Model setup

Define a FINN model and how each demographic process is represented.

finn(): Forest Informed Neural Network
createProcess(): Define a demographic process for FINN
createHybrid(): Define a hybrid (deep-neural-network) demographic process for FINN
CohortMat(): Cohort Matrix Class

Calibrate every component jointly, end-to-end, then simulate.

The built-in mechanistic demographic functions and allometries.

growth(): Calculate growth
mortality(): Mortality
regeneration(): Calculate the regeneration of forest patches based on the input parameters
competition(): Compute the fraction of available light (light) for each cohort based on the given parameters
height(): Calculate the height of a tree based on its diameter at breast height and an allometry parameter
BA_stand(): Calculate the Basal Area of a Stand
BA_stem(): Calculate the basal area of a tree given the diameter at breast height (dbh)
dbh2ba(): Convert DBH to basal area

Turn a raw tree list and environment table into FINN’s inputs.

makeObsData(): Create observation data from trees
resolveSiteIDs(): Resolve site, patch, and year indices for FINN inputs
makeInitCohorts(): Make initial cohorts for FINN
climateDF2array(): Convert a climate data frame to a FINN environment array
obsDF2arrays(): Convert observation data frame to arrays
array2obsDF(): Transform Arrays to Observation Data Table
pred2DF(): Convert Prediction Arrays to Data Frames
extract_env(): Extract Environmental Data for a Process

FINN standardises environmental predictors internally (env_autoscale = TRUE); these expose the stored constants.

apply_env_scaling(): Apply stored z-standardization to environmental predictors
compute_env_scaling(): Learn z-standardization for environmental predictors

Read back what a fitted - or hybrid - model actually learned.

ALE(): Accumulated local effect plots
plot(<FINNale>): Plot ALE curves of a FINN model
summary(<finn_class>): Summarise a fitted FINN model
conditionalEffects(): Conditional effects of a FINN model
averageConditionalEffects(): Average conditional effects of a FINN model
feature_importance(): Permutation feature importance for FINN demographic rates

Building blocks used internally and exported for advanced use.

aggregate_results_old(): Aggregate function
binomial_from_bernoulli(): Draw binomial counts from per-trial Bernoulli probabilities
binomial_from_gamma(): Sample from binomial with gradient
groupby_mean(): group by mean
index_species(): Index species
np_runif(): Generate random numbers from a uniform distribution
rweibull_cohorts(): Generate Cohorts Using Weibull Distribution
sample_poisson_relaxed(): Sample poisson relaxed