ComputeWeightsFromTree: Compute weights for 'TreeGuidedGroupLasso' from prior...

In tohein/linearMTL: Linear Models for Multi-Task Learning

Description

Generalize hierarchical clustering distances to compute node weights from the (response) matrix Y for a given grouping. Weights will be normalized using ComputeGroupWeights.

Usage

ComputeWeightsFromTree(Y, groups, linkage = "complete",
  distance = function(a, b) {     1 - cor(a, b) })

Arguments

Y	N by K matrix for every task.
groups	V by K matrix determining group membership: Task k in group v iff groups[v,k] == 1.
linkage	"complete" or "average". Linkage type used (distance measure between clusters).
distance	Function for computing pairwise distances between columns of Y. Default is 1-cor(a,b).

Value

Vector containing node weights.

See Also

TreeGuidedGroupLasso, ComputeGroupWeights.

tohein/linearMTL documentation built on May 17, 2019, 8:22 a.m.