Ward Hierarchical Clustering Method

Not used in group pipeline. Ward method minimizes the total within-cluster variance. At each step the pair of clusters with minimum cluster distance are merged. To implement this method, at each step find the pair of clusters that leads to minimum increase in total within-cluster variance after merging.

Description

In the ward linkage for each cluster a error function is defined. This error function is the average distance of each datapoint in a cluster to the center of gravity in the cluster.

D = Error function of unified cluster - Error function for each cluster

Parameters

kmax: Integer ( input )

The maximum number of clusters to consider

patch: Integer ( input )

The number corresponding to the gyrus

group_matrix: String ( input )

distance_matrix_file: String ( input )

average_mesh: White Mesh ( input )

Both average brain mesh (from a group of subject)

gyri_texture: ListOf( ROI Texture ) ( input )

List of all individual gyri segmentation (for example, bh.r.aparc.annot.gii of the FreeSurfer pipeline)

tex_time: ListOf( Connectivity ROI Texture ) ( output )

List of the clustering results, i.e. all individual parcellations of patch with a point-to-point correspodence across subject

Technical information

Toolbox : Constellation

User level : 2

Identifier : ward_method_group

File name : brainvisa/toolboxes/constellation/processes/group_pipeline/tools/ward_method_group.py

Supported file formats :

average_mesh :

GIFTI file, GIFTI file, MESH mesh, MNI OBJ mesh, PLY mesh, TRI mesh

gyri_texture :

GIFTI file, GIFTI file, Texture

tex_time :

GIFTI file, GIFTI file, Texture