polynomial.fit: Fit polynomial curves to species data using stepwise...

In ericgoolsby/phylocurve: Phylogenetic Comparative Methods for High-Dimensional Traits

Description

Transforms raw data into regression curves using stepwise (AIC or BIC) polynomial regression. The maximum number of parameters (nterms), response data can be constrained between minima and maxima (for example, the default sets any negative predicted y value to 0).

Usage

polynomial.fit(data, x_variable, y_variable, method = "BIC",
  nterms = 2, min_x = -Inf, max_x = Inf, min_y = 0, max_y = Inf, eval_length = 30)

Arguments

data	Data frame with function-valued species data along with a column named "species"
x_variable	The name of the column in data corresponding to the predictor (x) variable
y_variable	The name of the column in data corresponding to the response (y) variable
method	Metric for stepwise polynomial regression. Either "AIC" or "BIC" (the default).
nterms	Maximum number of terms to allow in polynomial regression.
min_x	Lower range to remove from raw x data
max_x	Upper range to remove from raw x data
min_y	Lower range to remove from predicted response (for example, for response data in which only positive values are allowed, the min_y can be set to 0 (the default)).
max_y	Upper range to remove from predicted response.
eval_length	How many points to evaluate along the range of x.

Value

X	X (formatted for phylocurve.generalized)
Y	Y (formatted for phylocurve.generalized)

Examples

x_length <- 50
nspecies <- 25
sim_data <- sim.curves(nspecies = nspecies,x_length = x_length)

# Fit polynomial curves to raw data
poly_data <- polynomial.fit(data = sim_data$data,x_variable = "x",y_variable = "y",min_y = 0)

# Align curve data using dynamic time warping
# NOT RUN (takes several minutes)
# P <- phylocurve.generalized(tree = sim_data$tree,X = poly_data$X,Y = poly_data$Y)

ericgoolsby/phylocurve documentation built on April 28, 2021, 10:23 p.m.