Search
 
loader loading subjects...
cover
A Scalable Approach to Econometric Inference 1
Philip Nadler, Rossella Arcucci, Yi-Ke Guo
59 - 68
10.3233/APC200025
Research Article
Advances in Parallel Computing
Volume 36: Parallel Computing: Technology Trends
Abstract

We propose a novel approach combining vector autoregressive models and data assimilation to conduct econometric inference for high dimensional problems in cryptocurrency markets. We label this new model TVP-VAR-DA. As the resulting algorithm is computationally very expensive, it mandates the introduction of a problem decomposition and its implementation in a parallel computing environment. We study its scalability and prediction accuracy under various specifications.

Download
Creative Commons License