gRapfa (version 1.0)

st: Sample tree

Description

Builds the sample tree of a discrete longitudinal dataset.

Usage

st(iD)

Arguments

iD

iD a data frame whose variables are factors. Any missing values are treated as additional factor levels.

Value

An igraph object containing the sample tree.

Details

Sample trees are constructed as follows. Suppose $N$ observations pf $p$ discrete variables $(x_1, ..., x_p)$ are given. Starting with $N$ observations at the root node, edges branches out to nodes at the first level. The number of branches corresponds to number of distinct values of $x_1$, and the count on each edge correspond to the frequency of occurrence of the respective value. From each node at level one, edges branch out to level two, based on the distinct values of $x_2$ given $x_1$. The process continues up to level $p$.

References

Ankinakatte, S. and Edwards, D. Modelling discrete longitudinal data using acyclic probabilistic finite automata. Submitted to C.S.D.A.

Examples

data(Wheeze)
G <- st(Wheeze)
E(G)$arrow.size <- 0.6
V(G)$size <- 10
V(G)$label <- ''
E(G)$label <- E(G)$count
plot(G)