symbolic_to_relprob_matrix_sp: Convert symbolic matrix (with text equations) to relprob matrix (numeric values) -- speciation matrix version

Description

This does the equivalent of symbolic_to_P_matrix, but for a speciation/cladogenesis matrix.

Usage

symbolic_to_relprob_matrix_sp(spmat, cellsplit = "\+", mergesym = "*", ys = 1, j = 0, v = 1, maxent_constraint_01 = 1e-04, maxent_constraint_01v = 1e-04, max_numareas = 6, ...)

Arguments

spmat

The speciation/cladogenesis matrix, with text formula.

cellsplit

The symbol to split the formulas on. Default "\+" (plus symbol, with escape code).

mergesym

The symbol to merge the formulas with. Default "+".

Relative weight of fully sympatric speciation (range-copying) and sympatric "subset" speciation. Default s=1 mimics LAGRANGE model.

Relative weight of vicariant speciation. Default v=1 mimics LAGRANGE model.

Relative weight of "founder event speciation"/jump speciation. Default j=0 mimics LAGRANGE model.

maxent_constraint_01

Parameter which assigns relative probabilities to different descendants range sizes, for the smaller descendant. Values can range from 0.0001 to 1. If maxent_constraint_01=0.0001, then the smaller descendant has a range size of 1 with probability 1 (i.e., the LAGRANGE default). If maxent_constraint_01=0.5, then all range sizes are equally weighted. If maxent_constraint_01=1, then the largest possible smaller descendant gets probability 1. The reference to "maxent" derives from the fact that the maxent probability distribution on a multistate, ordered, discrete variable -- e.g. a die roll -- can be calculated given just the mean value. Here, the maxent_constraint_01 parameter is multiplied by the (maximum rangesize + 1). Thus, when maxent_constraint_01=0.5, if there are 6 possible states, then the parameter becomes 3.5, which sets equal probabilities of all possible descendant ranges sizes, when range size can range from 1 to 6.

maxent_constraint_01v

Works the same as maxent_constraint_01, but just for descendants of vicariant events.

max_numareas

The maximum number of areas possible allowed for the smaller-ranged-daughter in either vicariant or sympatric types of cladogenesis/speciation.

...

Additional arguments to pass to relative_probabilities_of_subsets and relative_probabilities_of_vicariants, and thence to strsplit.

Value

cellval The output cell value.

Details

These are 1-event probability matrices, not instantaneous rate matrices.

This function uses symbolic_cell_to_relprob_cell_sp in an sapply call. It still will not be very fast compared to the calculations in cladoRcpp, but can be useful for demonstrative purposes.

References

http://phylo.wikidot.com/matzke-2013-international-biogeography-society-poster https://code.google.com/p/lagrange/

Matzke_2012_IBS

ReeSmith2008

Examples

Run this code

testval=1
# Generate the text version of the speciation/cladogenesis probability matrix
# (actually a relative weights matrix
# until the rows are normalized so that each sums to 1).
spmat = make_relprob_matrix_bi(states_list=list("_", c("A"), c("B"), c("C"),
c("A","B"), c("B","C"), c("A","C"), c("A","B","C")), split_ABC=FALSE, splitval="",
code_for_overlapping_subsets=NA, printwarn=1)
spmat

# Look at the conditional probabilities generated by a variety of models
spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+", mergesym="*",
ys=1, j=0, v=1, maxent_constraint_01=0.0001, maxent_constraint_01v=0.0001,
max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=0.5, j=0, v=0.5, maxent_constraint_01=0.0001,
maxent_constraint_01v=0.0001, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=1, j=1, v=1, maxent_constraint_01=0.0001,
maxent_constraint_01v=0.0001, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=0.25, j=0.25, v=0.25, maxent_constraint_01=0.0001,
maxent_constraint_01v=0.0001, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=1, j=1, v=0, maxent_constraint_01=0.0001,
maxent_constraint_01v=0.0001, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=1, j=1, v=0, maxent_constraint_01=0.5,
maxent_constraint_01v=0.0001, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=1, j=0, v=0, maxent_constraint_01=0.5,
maxent_constraint_01v=0.0001, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

spPmat = symbolic_to_relprob_matrix_sp(spmat=spmat, cellsplit="\\\\+",
mergesym="*", ys=1, j=0, v=1, maxent_constraint_01=0.0001,
maxent_constraint_01v=0.5, max_numareas=3)
spPmat = adf(spPmat); names(spPmat) = names(spmat); rownames(spPmat) = rownames(spmat)
spPmat

Run the code above in your browser using DataLab

Get 50% off unlimited learning