A0440
Title: Dynamic linear mixed models for time-dependent data analysis
Authors: Dario Ferreira - University of Beira Interior (Portugal) [presenting]
Sandra Ferreira - University of Beira Interior, Covilha (Portugal)
Patricia Antunes - University of Beira Interior (Portugal)
Gilberto Neves - Universidade Aberta Departamento de Ciencias e Tecnologia (Portugal)
Abstract: A time-varying linear mixed model (TVLMM), an innovative approach for analyzing time-dependent data, is introduced. Unlike traditional time-series models that assume constant random effects, TVLMM incorporates random effects that change over time, thereby providing a more realistic representation of real-world data dynamics. By integrating the predictive strengths of autoregressive integrated moving average (ARIMA) models with the flexibility of linear mixed models (LMMs), TVLMM addresses the limitations of conventional models in handling temporal variations in data. This approach is particularly relevant for fields such as finance, economics, social sciences, and biology, where the underlying data structures often evolve over time. The methodologies of ARIMA and LMMs are outlined, the parameter estimation process for TVLMM is detailed, and its application is illustrated through a numerical example. The results demonstrate the model's capability to produce accurate predictions by accounting for time-varying characteristics in random effects.
