Estimating Individual Dynamic Factor Models Using a Regularized Hybrid Unified Structural Equation Modeling with Latent Variable

Ye, Ai und Bollen, Kenneth A. (2023): Estimating Individual Dynamic Factor Models Using a Regularized Hybrid Unified Structural Equation Modeling with Latent Variable. 87th Annual Meeting of the Psychometric Society, Bologna, 2022. Wiberg, Marie; Molenaar, Dylan; González, Jorge; Kim, Jee-Seon und Hwang, Heungsun (Hrsg.): In: Quantitative Psychology, Cham: Springer. S. 325-334

Abstract

There has been an increasing call to model multivariate time series data with measurement error. The combination of latent factors with a vector autoregressive (VAR) model leads to the dynamic factor model (DFM), in which dynamic relations are derived within factor series, among factors and observed time series, or both. However, two limitations exist in the current DFM representatives and estimation: (1) the dynamic component of DFM contains either directed or undirected contemporaneous relations, but not both, and (2) selecting the optimal model in exploratory DFM is a challenge. Our paper serves to advance and evaluate DFM with hybrid VAR representations, which would then be estimated using LASSO regularization under the Structural Equation Model framework. This approach allows for the selection of the optimal hybrid dynamic relations in a data-driven manner. A simulation study is presented to investigate the sensitivity of finding the true hybrid dynamic relations in the structural model and the specificity of excluding the false relations using the LASSO-regularization versus the pseudo-ML approaches. We aim to offer guidance on model selection and estimation in person-centered dynamic assessments.

• Dokumententyp: Konferenzbeitrag (Paper)
• Keywords: Time series data; Hybrid unified SEM; Regularized SEM; Dynamic factor model; Model-implied instrumental variable two-stage least square
• Fakultät: Psychologie und Pädagogik > Department Psychologie > Psychologische Methodenlehre und Diagnostik
• Themengebiete: 100 Philosophie und Psychologie > 150 Psychologie
• Ort: Cham
• Sprache: Englisch