Ponente
Descripción
Score-driven (SD) models, also referred to as generalized autoregressive score (GAS) models and dynamic conditional score (DCS) models, are time series models that can be based on any underlying probability distribution with dynamics driven by the conditional score for any time-varying parameters. In recent years, score-driven models have emerged as a valuable modern methodology for time series modeling, especially in econometrics and quantitative finance. In this contribution, we introduce a novel R package, called gasmodel, that facilitates the estimation, forecasting, and simulation of score-driven models. We also present two applications related to operations research — modeling of dynamic rankings in data envelopment analysis (DEA) and modeling of inter-arrival times in queueing theory.
