lasec: Landmark Sampling Evaluation Curve

In akiopteryx/lambda: LandMark-Based Data Assessment

Description

Generate a sampling curve to assess whether sufficient number of landmarks has been sampled.

Usage

lasec(coord.data, n.dim, iter = 1000, show.progress = T, keep=NULL)

Arguments

coord.data	2-D or 3-D coordinate data arranged in n x p matrix with n=specimens, p=shape variables. NB: data should be unaligned for centroid size analysis).
n.dim	Number of spatial dimensions (i.e., 2 or 3 for 2-D or 3-D).
iter	Number of subsampling rounds.
keep	List of landmarks to keep in data during subsampling.

Details

The function reads in landmark data without any missing data and generates two sampling curves. "LaSEC_SamplingCurve_Shape.pdf" displays how sampling additional landmarks converges to the distribution of shape variation in the parent dataset. "LaSEC_SamplingCurve_CS.pdf" displays how sampling additional landmarks converges to the distribution of centroid size variation in the parent dataset. These sampling curves allows the user to determine whether landmark sampling is adequate to capture the shape and size variation of specimens by looking at the extent of the plateau in the curve and constricting fit (1-PSS) values.

Value

Function returns the following:

fit	A matrix of 1–PSS values where each row comprises sequence of fit values at each subsampling of landmarks with respect to shape variation
median.fit	A vector comprising median 1–PSS values at each subsampling of landmarks
fit.cs	A matrix of 1–PSS values where each row comprises sequence of fit values at each subsampling of landmarks with respect to centroid size variation

Author(s)

Akinobu Watanabe

References

Watanabe, A. 2018. How many landmarks are enough to characterize shape and size variation? PLoS ONE 13(6): e0198341. https://doi.org/10.1371/journal.pone.0198341

Examples

# Produces landmark sampling curves automatically:
data(crocskull)
lasec(coord.data=crocskull, n.dim=3, iter=10)
# Plot landmark sampling curves manually:
data(crocskull)
lasec.results <- lasec(coord.data=crocskull, n.dim=3, iter=10)
plot.lasec(lasec.results)

akiopteryx/lambda documentation built on Dec. 26, 2020, 7:36 p.m.