Flat representation of a probabilistic suffix tree

Share:

Description

The class "PSTf" is the flat representation of a probabilistic suffix tree (PST) storing a variable length Markov chain model. The flat representation is a list where each element corresponds to a given depth. It is the prefered representation and is used by all functions for model fitting and sequence analysis with PST. The nested representation "PSTr" is used only for printing and plotting PSTs.

Objects from the Class

Objects of class "PSTf" are returned by the pstree, prune and tune function.

Slots

.Data:

Object of class "list", a list where each element corresponds to one level of the tree and is itself a list of nodes, i.e., objects of class "PSTr".

data:

Object of class "stslist". The learning sample to which the PST is fitted, i.e., a sequence object created with the seqdef function.

cdata:

Object of class "stslist"

alphabet:

Object of class "character". Alphabet on which the sequences, and the PST are built.

labels:

Object of class "character" containing the long state labels.

cpal:

Object of class "character". Color palette used to represent each state of the alphabet.

segmented:

Object of class "logical" indicating whether the tree is segmented. See pstree.

group:

Object of class "factor" containing the group membership for each sequence in data.

call:

Object of class "call".

logLik:

Object of class "numeric", containing the log-likelihood of the VLMC model represented by the PST.

Extends

Class "list", from data part. Class "vector", by class "list", distance 2.

Methods

cmine

signature(object = "PSTf"): context mining, see cmine,PSTf-method.

cplot

signature(object = "PSTf"): plot single nodes of a PST, see cplot,PSTf-method.

generate

signature(object = "PSTf"): generate artificial sequences, see generate,PSTf-method.

impute

signature(object = "PSTf", data = "stslist"): impute missing values in sequence data, seeimpute,PSTf,stslist-method.

logLik

signature(object = "PSTf"): extract log-likelihood of the VLMC model represented by a PST, see logLik,PSTf-method.

nobs

signature(object = "PSTf"): number of observations (symbols) in the learning sample to which a VLMC model is fitted, see nobs,PSTf-method.

nodenames

signature(object = "PSTf"): retrieve the node labels of a PST, see see nodenames,PSTf-method.

pdist

signature(x = "PSTf", y = "PSTf"): compute probabilistic divergence between two PSTs, see pdist,PSTf,PSTf-method.

plot

signature(x = "PSTf", y = "ANY"): plot a PST, see plot,PSTf,ANY-method.

pmine

signature(object = "PSTf", data = "stslist"): pattern mining, see see pmine,PSTf,stslist-method.

ppplot

signature(object = "PSTf"): plotting a branch of a PST, see ppplot,PSTf-method.

pqplot

signature(object = "PSTf", data = "stslist"): plot the predicted probability of each state in a sequence, see pqplot,PSTf,stslist-method.

predict

signature(object = "PSTf"): predict the likelihood of sequences, see predict,PSTf-method.

print

signature(x = "PSTf"): print a PST, see print,PSTf-method.

prune

signature(object = "PSTf"): prune a PST, see prune,PSTf-method.

query

signature(object = "PSTf"): retrieve counts or next symbol probability distribution from a node in a Probabilistic Suffix Tree, see query,PSTf-method.

subtree

signature(object = "PSTf"): extract a subtree from a segmented PST, see subtree,PSTf-method.

summary

signature(object = "PSTf"): see summary,PSTf-method.

tune

signature(object = "PSTf"): AIC, AICc and BIC based model selection, see tune,PSTf-method.

Author(s)

Alexis Gabadinho

See Also

PSTr

Examples

1
showClass("PSTf")