march.mtd.construct {march} | R Documentation |
A Mixture Transition Distribution model (march.Mtd-class
) object of order order is constructed
according to a given march.Dataset-class
y. The first maxOrder-order
elements of each sequence are truncated in order to return a model
which can be compared with other Markovian models of visible order maxOrder.
march.mtd.construct( y, order, maxOrder = order, mtdg = FALSE, MCovar = 0, init = "best", deltaStop = 1e-04, llStop = 0.01, maxIter = 0, seedModel = NULL )
y |
the dataset ( |
order |
the order of the constructed model. |
maxOrder |
the maximum visible order among the set of Markovian models to compare. |
mtdg |
flag indicating whether the constructed model should be a MTDg using a different transition matrix for each lag (value: TRUE or FALSE). |
MCovar |
vector of the size Ncov indicating which covariables are used (0: no, 1:yes) |
init |
the init method, to choose among best, random and weighted. |
deltaStop |
the delta below which the optimization phases of phi and Q stop. |
llStop |
the ll increase below which the EM algorithm stop. |
maxIter |
the maximal number of iterations of the optimisation algorithm (zero for no maximal number). |
seedModel |
an object containing a MTD or a DCMM model used to initialize the parameters of the MTD model. |
Ogier Maitre, Kevin Emery, Andre Berchtold
march.Mtd-class
, march.Model-class
, march.Dataset-class
.
# Build a 4th order MTD model from the pewee data set. model <- march.mtd.construct(pewee,4) print(model) # Build a 3th order MTDg model from the pewee data set. model <- march.mtd.construct(pewee,3,mtdg=TRUE) print(model)