Risk Analysis in Big Data

7 Pages Posted: 11 Aug 2015

See all articles by Jonathan Bakdash

Jonathan Bakdash

US Army Research Laboratory

Laura Marusich

US Army Research Laboratory

Date Written: August 10, 2015

Abstract

Society has become increasingly interconnected with networks linking people, the environment, information, and technology. This complexity is a challenge for risk analysis. Traditional risk analysis methods typically underestimate the probability and impact of risks (e.g., terrorist attacks, power failures, and natural disasters such as hurricanes) because normal data and independent observations are assumed. Cascading failures occur in complex systems, such as a hurricane that causes a power failure resulting in flooding. Big data offers enormous promise for improving risk analysis, but cannot replace the importance of appropriate assumptions, data quality, continued validation, and human understanding of risk analysis for effective risk mitigation. The blessing and curse of big data for risk analysis is illustrated by the example of Google Flu Trends, which was accurate at first but then became inaccurate.

Keywords: Risk analysis, risk, big data, decision making, dynamical systems

JEL Classification: D81, D80, C80, E17

Suggested Citation

Bakdash, Jonathan and Marusich, Laura, Risk Analysis in Big Data (August 10, 2015). Available at SSRN: https://ssrn.com/abstract=2641726 or http://dx.doi.org/10.2139/ssrn.2641726

Jonathan Bakdash (Contact Author)

US Army Research Laboratory ( email )

ATTN: RDRL-HRS-E
459 Mulberry Point Road
Aberdeen Proving Ground, MD 21005-5425
United States

Laura Marusich

US Army Research Laboratory ( email )

ATTN: RDRL-HRS-E
459 Mulberry Point Road
Aberdeen Proving Ground, MD 21005-5425
United States

Register to save articles to
your library

Register

Paper statistics

Downloads
250
rank
119,930
Abstract Views
1,226
PlumX Metrics