hierarchical_matrix: Create Hierarchical Structure Matrix

In yForecasting/leadingindicators: Find relevant leading indicators for time series forecasting

Description

Function that creates a matrix that represents the hierarchial structure of a dataset. In a symmetric hierarchy with 3 levels where the top level is the sum of all, and the lower level are all individual time series, this argument can be set to 2, while only specifying the mid level nodes in the previous argument.

Usage

hierarchical_matrix(nodes, nodeshier)

Arguments

nodes	Vector of number of time series for each hierarchical level. The lowest level (all ones) and the top level (sum of all) should not be included.
nodeshier	Vector that corresponds to the vector of nodes, describing on which level each node is. The lowest number in this vector is 2, being the node level above the lowest hierarchical time series. When this argument is empty, a 3-level hierarchy is assumed.

Examples

hierarchical_matrix(nodes=c(1,2), nodeshier=c(2,2))

hierarchical_matrix(nodes=c(3,5,1,2,3,2), nodeshier=c(3,3,2,2,2,2))

yForecasting/leadingindicators documentation built on Dec. 12, 2021, 3:09 p.m.