Equivalence Set Restricted Latent Class Models (ESRLCM)

TL;DR

The paper introduces ESRLCM, a Bayesian clustering model for categorical data that generalizes traditional latent class models by identifying clusters with shared response probabilities, with proven identifiability and demonstrated effectiveness.

Contribution

It proposes a novel ESRLCM that extends latent class models to identify clusters with common response probabilities more flexibly than existing models.

Findings

Model is identifiable under certain conditions

Effective in simulations and real data

Outperforms traditional models in clustering accuracy

Abstract

Latent Class Models (LCMs) are used to cluster multivariate categorical data, commonly used to interpret survey responses. We propose a novel Bayesian model called the Equivalence Set Restricted Latent Class Model (ESRLCM). This model identifies clusters who have common item response probabilities, and does so more generically than traditional restricted latent attribute models. We verify the identifiability of ESRLCMs, and demonstrate the effectiveness in both simulations and real-world applications.