Home // SIGNAL 2022, The Seventh International Conference on Advances in Signal, Image and Video Processing // View article

Discovering Causality in Event Time-Series

Authors:
Pavel Loskot

Keywords: causal; dynamic system; event; matrix profile; state- space; time-series

Abstract:
The event time-series can accurately describe the behavior of many dynamic systems. The challenge is that the events are categorical variables, so they cannot be analyzed by the existing statistical methods developed for numerical time-series. In order to infer the causally related events, in this paper, it is proposed to assume the empirical conditional probabilities of nearly certain and nearly uncertain events. Moreover, since the event ordering is usually locally irrelevant, the event sequences can be transformed into the event sets or multi-sets with appropriately defined distance metrics. The event sequences having a zero distance can be then assumed to be causally equivalent. The distance metrics are also used in matrix profile analysis of event time-series. Numerical examples are studied for chemical reaction events generated in stochastic simulations of biochemical molecular systems. Even though the proposed framework for discovering the causally related event sequences can be readily fully automated, they still need to be properly interpreted in the context of relevant domain knowledge.

Pages: 18 to 23

Copyright: Copyright (c) IARIA, 2022

Publication date: May 22, 2022

Published in: conference

ISSN: 2519-8432

ISBN: 978-1-61208-970-6

Location: Venice, Italy

Dates: from May 22, 2022 to May 26, 2022