find_best_split

Internal function to iterate through all possible split points in all
covariates to find the split that minimizes the split statistic.

internal

Implements the Factor-Augmented Clustering Tree (FACT) algorithm
for clustering time series data. The method constructs a classification
tree where splits are determined by covariates, and the splitting criterion
is based on a group factor model representation of the time series within
each node. Both threshold-based and permutation-based tests are supported
for splitting decisions, with an option for parallel computation.
For methodological details, see Hu, Li, Luo, and Wang (2025, in preparation),
Factor-Augmented Clustering Tree for Time Series.

Jiaqi Hu

factree

Factor-Augmented Clustering Tree

Ting Li

Zidan Luo

Xueqin Wang

find_best_split function

<dl><dt>X_node</dt>
<dd>Covariate matrix for the current node.</dd>
<dt>Y_node</dt>
<dd>Time series matrix for the current node.</dd>
<dt>indices</dt>
<dd>Original indices of the observations in the current node.</dd>
<dt>r_a</dt>
<dd>Parameter passed to `calculate_split_statistic`.</dd>
<dt>r_b</dt>
<dd>Parameter passed to `calculate_split_statistic`.</dd>
<dt>control</dt>
<dd>The parameters used in split.</dd></dl>

Arguments

Find the Best Split for a Node — find_best_split

<dl>

<dt>X_node</dt>
<dd>Covariate matrix for the current node.</dd>


<dt>Y_node</dt>
<dd>Time series matrix for the current node.</dd>


<dt>indices</dt>
<dd>Original indices of the observations in the current node.</dd>


<dt>r_a</dt>
<dd>Parameter passed to `calculate_split_statistic`.</dd>


<dt>r_b</dt>
<dd>Parameter passed to `calculate_split_statistic`.</dd>


<dt>control</dt>
<dd>The parameters used in split.</dd>

</dl>

find_best_split: Find the Best Split for a Node

Description

Usage

Value

Arguments