Description Usage Arguments Details Value Author(s) See Also Examples
The function runs EXPOKIT's dgexpv
function on a Q
matrix and one or more time values. If
Qmat
is NULL (default), a default matrix is
input.
1 2 3 4 
Qmat 
an input Q transition matrix 
tvals 
one or more time values to exponentiate by (doesn't have to literally be a time value, obviously) 
inputprobs_for_fast 
If NULL (default), the full
probability matrix (Pmat) is returned. However, the full
speed of EXPOKIT on sparse matrices will be exploited if
inputprobs_for_fast=c(starting probabilities). In this
case these starting probabilities are input to

transpose_needed 
If TRUE (default), matrix will be transposed (apparently EXPOKIT needs the input matrix to be transposed compared to normal) 
transform_to_coo_TF 
Should the matrix be tranposed
to COO? COO format is required for EXPOKIT's
sparsematrix functions (like dmexpv and unlike the
padmrelated functions. Default TRUE; if FALSE, user must
put a COOformated matrix in 
coo_n 
If a COO matrix is input, 
force_list_if_1_tval 
Default FALSE, but set to TRUE if you want a single matrix to be returned inside a list 
check_for_0_rows 
If TRUE or a numeric value, the input Qmat is checked for allzero rows, since these will crash the FORTRAN wrapalldmexpv function. A small nonzero value set to check_for_0_rows or the default (0.0000000000001) is input to offdiagonal cells in the row (and the diagonal value is normalized), which should fix the problem. 
NOTE: DGEXPV vs. DMEXPV. According to the EXPOKIT
documentation, DGEXPV should be faster than DMEXPV,
however DMEXPV runs an accuracy check appropriate for
Markov chains, which is not done in DGEXPV.
tmpoutmat
the output matrix, if 1 tvalue is
input; list_of_matrices_output
, if more than 1
tvalue is input; to get a single output matrix in a
list, set force_list_if_1_tval=TRUE
Nicholas J. Matzke nickmatzke.ncse@gmail.com and Drew Schmidt schmidt@math.utk.edu
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39  # Example:
# Make a square instantaneous rate matrix (Q matrix)
# This matrix is taken from Peter Foster's (2001) "The Idiot's Guide
# to the Zen of Likelihood in a Nutshell in Seven Days for Dummies,
# Unleashed" at:
# \url{http://www.bioinf.org/molsys/data/idiots.pdf}
#
# The Q matrix includes the stationary base freqencies, which Pmat
# converges to as t becomes large.
Qmat = matrix(c(1.218, 0.504, 0.336, 0.378, 0.126, 0.882, 0.252, 0.504, 0.168,
0.504, 1.05, 0.378, 0.126, 0.672, 0.252, 1.05), nrow=4, byrow=TRUE)
# Make a series of t values
tvals = c(0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1, 2, 5, 14)
# Exponentiate each with EXPOKIT's dgexpv (should be fast for large sparse matrices)
for (t in tvals)
{
Pmat = expokit_dgexpv_Qmat(Qmat=Qmat, t=t, transpose_needed=TRUE)
cat("\n\nTime=", t, "\n", sep="")
print(Pmat)
}
# DMEXPV and DGEXPV are designed for large, sparse Q matrices (sparse = lots of zeros).
# DMEXPV is specifically designed for Markov chains and so may be slower, but more accurate.
# DMEXPV, single tvalue
# DGEXPV, single tvalue
expokit_wrapalldgexpv_tvals(Qmat=Qmat, tvals=tvals[1], transpose_needed=TRUE)
expokit_wrapalldgexpv_tvals(Qmat=Qmat, tvals=2)
# These functions runs the forloop itself (sadly, we could not get mapply() to work
# on a function that calls dmexpv/dgexpv), returning a list of probability matrices.
# DGEXPV functions
list_of_P_matrices_dgexpv = expokit_wrapalldgexpv_tvals(Qmat=Qmat,
tvals=tvals, transpose_needed=TRUE)
list_of_P_matrices_dgexpv

Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.