XLSTAT: Choosing a Clustering Method in Excel

Choose AHC for dendrograms, k‑means for fixed k, Gaussian mixtures or Latent Class for probabilistic membership; use PCA/MCA scores to cluster categorical data.

Jul 29, 2026
Knowledge

Title

XLSTAT: Choosing a Clustering Method in Excel

Answer

XLSTAT: Choosing a Clustering Method in Excel

Background

Cluster analysis groups objects into classes based on similarity across measured variables. It's widely used in:

Omics data: to find similar expression profiles.
Market research: to identify consumer segments.
Ecology: to group similar sites by community type.

Choosing the Best Method

XLSTAT offers several clustering methods for quantitative data:

Method	Variables Needed	Input Type	Class Selection	Membership Type	Special Features
AHC	≥1	Quantitative (or binary)	Optional	Deterministic	Dendrogram, profile plot
k-means	≥1	Quantitative	Mandatory	Deterministic	Profile plot
Gaussian Mixture	≥1	Quantitative	Mandatory	Probabilistic	Class parameters, MAP classification plot
Univariate Clustering	1	Quantitative	Mandatory	Deterministic	–
Latent Class (XLSTAT-LG)	≥1	Quant./Ordinal/Nominal	Mandatory	Probabilistic	Variable contributions, scoring for new data

For qualitative variables, first run Multiple or Correspondence Analysis, then cluster using the observation scores.

Cluster membership can be:

Deterministic: each observation belongs to one class.
Probabilistic: shows likelihood of belonging to each class.

For large datasets, methods can be combined—for example, use k-means to form initial clusters, then apply AHC to those cluster centers.

URL Name

6719-which-clustering-method-should-you-choose