Dependent Latent Class Models

TL;DR

This paper introduces Dependent Latent Class Models (DLCMs), a Bayesian extension of traditional LCMs that allows for conditional dependence, improving clustering accuracy in complex categorical data scenarios.

Contribution

The paper develops and verifies a novel Bayesian DLCM that relaxes the conditional independence assumption of traditional LCMs, enhancing modeling flexibility.

Findings

DLCMs are effective in time series data.

DLCMs handle overlapping items better.

DLCMs manage structural zeroes successfully.

Abstract

Latent Class Models (LCMs) are used to cluster multivariate categorical data (e.g. group participants based on survey responses). Traditional LCMs assume a property called conditional independence. This assumption can be restrictive, leading to model misspecification and overparameterization. To combat this problem, we developed a novel Bayesian model called a Dependent Latent Class Model (DLCM), which permits conditional dependence. We verify identifiability of DLCMs. We also demonstrate the effectiveness of DLCMs in both simulations and real-world applications. Compared to traditional LCMs, DLCMs are effective in applications with time series, overlapping items, and structural zeroes.