04-dtree: Create a Decision tree

In Condens8R: Classes and Methods for Learning Complex Decision Tree Models

Create a Decision tree

Description

Functions to create a binary decision tree with a complex algorithm at each nternal node.

Usage

createTree(data, metric, label = "H", pcut = 0.05, N = 100, splitter =
"hc", modeler = "logic")
registerPredictor(tag, description, modeler)
availablePredictors()

Arguments

data	A data matrix where (numerical) features are rows and samples are columns.
metric	A distance metric. This metric should be one of the values supported by either the distanceMatrix function from the ClassDiscovery package or the binaryDistance function from the Mercator package. All metrics computed by the dist function are included.
label	Character use to mark the root note. (default is "H" for "head".)
pcut	Significance cutoff ("alpha") on the empirical p-value.
splitter	A tag that uniquely identifies one of the splitting functions that have been registered. Built-in registered "splitters", with their tags, include: (hc) agglomeratve hierarchical clustering; (km) k-means clustering; (dv) divisive heirarchical clustering, and (ap) affinity propagation clustering.
modeler	A tag that uniquely identifies one of the predictor algorithms that have been registered. Built-in registered "predictors", with their tags, include: (logic) logic regression; (svm) support vector machines; (lr) logitic regression, (rf) random forests, and (tr) tail rank.
N	Number of random splits used to estimate emoirical p-value.
tag	A character string representing a short abbreviation used to identify and eventually call the associated "predictor" method that can dichtomize the data.
description	A character string containing the full name of the prediction algorithm. Potentially, this can be used in data summaries or plots.

Details

The core idea of this package is to construct decision trees that simulataneouly learn how to cluster the items in a data set while also learning a classifier that can predict the cluster assignments of new data. We support four built-in clustering algortihms to learn how to split a single data set into two parts:

hc

agglomerative hierarchical clustering as implemented in the hclust function, followed by cutree.

km

kmeans clustering with k = 2.

dv

divisive hierarchical clustering as implemented in the diana function.

ap

affinity propagation clustering, as implemented in the apcluster function.

We also support multiple classification algorithms to learna model to predict the binary split at the node:

logic

A modeler that implemts the "logic regression" approach to combine binary predictprs using logical operations, as implemented in the logicFS package.

svm

Support vector machines.

lr

The traditional logistic regression approach.

rf

Random forests.

tr

The "tail-rank" algorithm.

To get a list of the available methods, use the functions availableSplitters and availablePredictors. You can add your own algorithm with the functions registerSplitter pr registerPredictor. A "splitter" should accept a distance matrix as input and return a clustering vector with labels "1" or "2" for the classes. A "predictpr", by contrast, should be an object of the Modeler-class.

Value

Returns a BinaryNode object, which includes a

Author(s)

Kevin R. Coombes <[e-mail]>

Examples

nr <- 200 # features
nc <- 60  # samples
set.seed(80123)
comat <- matrix(rnorm(nr*nc, 0, 1), nrow = nr)
dimnames(comat) <- list(paste0("F", 1:nr),
                        paste0("S", 1:nc))
splay <- rep(c(2, -2), each = nc/2)
for(J in 1:30) comat[J, ] <- comat[J, ] + splay
C.hc <- findSplit(comat, splitter = "hc")
evalSplit(C.hc, comat, "euclid", "sw")

Condens8R documentation built on May 28, 2025, 3 a.m.