Clustering
Overview
Clustering in BayesiaLab groups records or nodes to reveal structure that is not directly observed in the data. Data Clustering partitions the records described by a set of nodes, while Variable Clustering groups the nodes themselves into classes of related variables.
Multiple Clustering is one of the steps of the Probabilistic Structural Equation Model (PSEM) workflow. It iteratively applies Data Clustering on subsets of data defined by the classes of variables to create latent, or factor, nodes that represent the hidden causes sensed by the manifest nodes.