EcoSta 2021: Start Registration
View Submission - EcoSta2021
A0223
Title: Modified upper prediction limit of vector autoregressive model: the case for estimating value-at-risk Authors:  Aniq Rohmawati - Telkom University (Indonesia) [presenting]
Abstract: An important issue in a modern society concerning on a stability system is people entrust risk assessment as safeguard. This paper proposes the appropriate modelling for upper limit prediction corresponding to risk/loss for future observation. In most events, each observation inherently allows a causal relationship to other observations, possibly as a linear function. A Vector Autoregressive (VAR) model handling the instantaneous interaction between response and predictors over time series horizon is proposed. It is further recognized that VAR allowed lagged values of multivariate time series, also outlined time-varying parameter to address essential drifts in coefficient. We shall examine a measure of upper prediction namely modified Value-at-Risk, addressed by involving parameters estimation of VAR model. Result shows that VAR model allows to predict accurately the coverage probability of Value-at-Risk.