Toward efficient online estimation of dynamic structural equation models: a case study

Model parameter estimation from data is a substantive part of statistical modeling, including dynamic structural equation modeling. With the advent of the big data era, which makes large amounts of data accessible near real-time, online parameter estimation has become more crucial than ever before. Nevertheless, the dynamic structural equation modeling field remains somewhat limited in this regard. In response to these circumstances, the article takes a step toward efficient online estimation in the dynamic structural equation modeling context by developing one such algorithm through raw-data maximum likelihood estimation and a recursive (incremental) approach for a specific dynamic structural equation model encompassing the noisy Gaussian random walk model with input from the factor-analytic model. The proposed algorithm is then verified through Monte Carlo simulation. Even though the algorithm serves only a special case of the general dynamic structural equation model, this case study can be considered a basis for further research in a bottom-up manner.