Modelos de transición de clases latentes
Date
2024-10-31
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Pontificia Universidad Católica del Perú
Abstract
A pesar de que los modelos de clases latentes han recibido gran atención en la evaluación
social y otros campos, estos modelos son estáticos o de corte transversal y hay poco trabajo
investigativo en modelos longitudinales. Este trabajo estudia el análisis de transición de clases
latentes (LTA), observando el cambio en la clasificación de clases a través del tiempo. Se
desarrolla el modelo teórico para dos períodos y se extiende luego este para más de dos
períodos, estudiándose sus parámetros y estimación a través del algoritmo de Esperanza-
Maximización (EM). Se detallan también criterios para la selección de modelos definiendo
teóricamente la razón de su uso. Finalmente, se realiza un estudio de aplicación en Mplus y
en R sobre una base de datos de acoso escolar(bullying).
Although latent class models have received great attention in social assessment and other fields, these models are static or cross-sectional and there is little research work on longitudinal models. This work studies latent class transition analysis (LTA), observing the change in class classification over time. The theoretical model is developed for two periods and then extended for more than two periods, studying its parameters and estimation through the Expectation- Maximization (EM) algorithm. Criteria for selecting models are also detailed, theoretically defining the reason for their use. Finally, an application study is carried out in Mplus and R on a bullying database.
Although latent class models have received great attention in social assessment and other fields, these models are static or cross-sectional and there is little research work on longitudinal models. This work studies latent class transition analysis (LTA), observing the change in class classification over time. The theoretical model is developed for two periods and then extended for more than two periods, studying its parameters and estimation through the Expectation- Maximization (EM) algorithm. Criteria for selecting models are also detailed, theoretically defining the reason for their use. Finally, an application study is carried out in Mplus and R on a bullying database.
Description
Keywords
Análisis de estructura latente, Algoritmos--Probabilidad-maximizacion, Invarianza (Matemáticas)
Citation
Collections
Endorsement
Review
Supplemented By
Referenced By
Creative Commons license
Except where otherwised noted, this item's license is described as info:eu-repo/semantics/openAccess