HTD-DAG: HTD-DAG

Description

Implementation of a top-down procedure to correct the scores of the hierarchy according to the constraints that the score of a node cannot be greater than a score of its parents.

Usage

htd(S, g, root = "00")

Arguments

a named flat scores matrix with examples on rows and classes on columns.

a graph of class graphNEL. It represents the hierarchy of the classes.

root

name of the class that it is the top-level (root) of the hierarchy (def:00).

Value

a matrix with the scores of the classes corrected according to the HTD-DAG algorithm.

Details

The HTD-DAG algorithm modifies the flat scores according to the hierarchy of a DAG through a unique run across the nodes of the graph. For a given example $x \in X$, the flat predictions $f(x) = \hat{y}$ are hierarchically corrected to $\bar{y}$, by per-level visiting the nodes of the DAG from top to bottom according to the following simple rule: $$ \bar{y}_i := \left\{ \begin{array}{lll} \hat{y}_i & {\rm if} \quad i \in root(G) \\ \min_{j \in par(i)} \bar{y}_j & {\rm if} \quad \min_{j \in par(i)} \bar{y}_j < \hat{y}_i \\ \hat{y}_i & {\rm otherwise} \end{array} \right. $$ The node levels correspond to their maximum path length from the root.

Examples

Run this code

# NOT RUN {
data(graph);
data(scores);
root <- root.node(g);
S.htd <- htd(S,g,root);
# }