Model Selection Information Criteria for Non-Nested Latent Class Models

1997 ◽  
Vol 22 (3) ◽  
pp. 249-264 ◽  
Author(s):  
Ting Hsiang Lin ◽  
C. Mitchell Dayton
Keyword(s):  
Model Selection ◽  
Latent Class ◽  
Error Rates ◽  
Latent Class Models ◽  
Information Criteria ◽  
Behavioral Analysis ◽  
Nested Models ◽  
Complex Models ◽  
Class Models ◽  
Simple Models

Latent class models have been developed for assessment of hierarchic relations in scaling and behavioral analysis. This article investigated the use of three model selection information criteria—Akaike AIC, Schwarz SIC, and Bozdogan CAIC—for non-nested models. In general, SIC and CAIC were superior to AIC for relatively simple models, whereas AIC was superior for more complex models, although accuracy was often quite low for such models. In addition, some effects were detected for error rates in the models.

A Two-Stage Conditional Estimation Procedure for Unrestricted Latent Class Models

1980 ◽  
Vol 5 (2) ◽  
pp. 129-156 ◽  
Author(s):  
George B. Macready ◽  
C. Mitchell Dayton
Keyword(s):  
General Class ◽  
Latent Class ◽  
Estimation Procedure ◽  
Error Rates ◽  
Latent Class Models ◽  
Parameter Estimates ◽  
Probability Models ◽  
Two Stage ◽  
Conditional Estimation ◽  
Class Models

A variety of latent class models has been presented during the last 10 years which are restricted forms of a more general class of probability models. Each of these models involves an a priori dependency structure among a set of dichotomously scored tasks that define latent class response patterns across the tasks. In turn, the probabilities related to these latent class patterns along with a set of “Omission” and “intrusion” error rates for each task are the parameters used in defining models within this general class. One problem in using these models is that the defining parameters for a specific model may not be “identifiable.” To deal with this problem, researchers have considered curtailing the form of the model of interest by placing restrictions on the defining parameters. The purpose of this paper is to describe a two-stage conditional estimation procedure which results in reasonable estimates of specific models even though they may be nonidentifiable. This procedure involves the following stages: (a) establishment of initial parameter estimates and (b) step-wise maximum likelihood solutions for latent class probabilities and classification errors with iteration of this process until stable parameter estimates across successive iterations are obtained.

Mastery Assessment with Latent Class and Quasi-Independence Models Representing Homogeneous Item Domains

1980 ◽  
Vol 5 (1) ◽  
pp. 65-81 ◽  
Author(s):  
John R. Bergan ◽  
Anthony A. Cancelli ◽  
John W. Luiten
Keyword(s):  
Latent Class ◽  
Decision Rules ◽  
Latent Class Models ◽  
Latent Classes ◽  
Classification Techniques ◽  
Complex Models ◽  
Class Models

This article discusses mastery classification involving the use of latent class and quasi-independence models. Extensions of mastery classification techniques developed by Macready and Dayton are presented. These extensions provide decision rules for assigning individuals to latent classes in complex models involving more than two latent categories. Procedures for identifying the minimally acceptable proportion of misclassified individuals in complex latent class models are also detailed.

scholarly journals Effective Dimensions of Hierarchical Latent Class Models

2004 ◽  
Vol 21 ◽  
pp. 1-17 ◽  
Author(s):  
N. L. Zhang ◽  
T. Kocka
Keyword(s):  
Model Selection ◽  
Bayesian Networks ◽  
Latent Class ◽  
Empirical Studies ◽  
Latent Class Models ◽  
Penalty Term ◽  
Tree Models ◽  
Effective Dimensions ◽  
Class Models ◽  
Model Dimension

Hierarchical latent class (HLC) models are tree-structured Bayesian networks where leaf nodes are observed while internal nodes are latent. There are no theoretically well justified model selection criteria for HLC models in particular and Bayesian networks with latent nodes in general. Nonetheless, empirical studies suggest that the BIC score is a reasonable criterion to use in practice for learning HLC models. Empirical studies also suggest that sometimes model selection can be improved if standard model dimension is replaced with effective model dimension in the penalty term of the BIC score. Effective dimensions are difficult to compute. In this paper, we prove a theorem that relates the effective dimension of an HLC model to the effective dimensions of a number of latent class models. The theorem makes it computationally feasible to compute the effective dimensions of large HLC models. The theorem can also be used to compute the effective dimensions of general tree models.

On the robustness of latent class models for diagnostic testing with no gold standard

2021 ◽  
Author(s):  
Matthew R. Schofield ◽  
Michael J. Maze ◽  
John A. Crump ◽  
Matthew P. Rubach ◽  
Renee Galloway ◽  
...  
Keyword(s):  
Gold Standard ◽  
Latent Class ◽  
Diagnostic Testing ◽  
Latent Class Models ◽  
No Gold Standard ◽  
Class Models
Sign in / Sign up

Export Citation Format

Share Document