Department of Computer Science and Engineering
Persistent link for this community
Browse
Browsing Department of Computer Science and Engineering by Author "Agrawal, Saurabh"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Finding Novel Multivariate Relationships in Time Series Data: Applications to Climate and Neuroscience(2018-02-12) Agrawal, Saurabh; Steinbach, Michael; Boley, Daniel; Liess, Stefan; Chatterjee, Snigdhansu; Kumar, Vipin; Atluri, GowthamIn many domains, there is significant interest in capturing novel relationships between time series that represent activities recorded at different nodes of a highly complex system. In this paper, we introduce multipoles, a novel class of linear relationships between more than two time series. A multipole is a set of time series that have strong linear dependence among themselves, with the requirement that each time series makes a significant contribution to the linear dependence. We demonstrate that most interesting multipoles can be identified as cliques of negative correlations in a correlation network. Such cliques are typically rare in a real-world correlation network, which allows us to find almost all multipoles efficiently using a clique-enumeration approach. Using our proposed framework, we demonstrate the utility of multipoles in discovering new physical phenomena in two scientific domains: climate science and neuroscience. In particular, we discovered several multipole relationships that are reproducible in multiple other independent datasets, and lead to novel domain insights.Item Tripoles: A New Class of Climate Teleconnections(2015-12-11) Agrawal, Saurabh; Atluri, Gowtham; Liess, Stefan; Chatterjee, Snigdhansu; Kumar, VipinTeleconnections in climate represent a persistent and large-scale temporal connection in a given climate variable between two distant geographical regions. They are known to impact and explain the variability in climate of many regions across the globe and have been a subject of interest to climatologists. Traditionally, climate teleconnections have been studied as a persistent relationship between a pair of geographical regions (e.g. North Atlantic Oscillation (NAO), and El-Nino Southern Oscillation (ENSO)). In this report, we define a new class of climate teleconnections which we refer to as tripoles that capture climatic relationships between three regions, in contrast to teleconnections that are traditionally defined using only two regions. We further provide a categorization of tripoles based on pairwise relationships between the three participating regions and propose a shared nearest neighbor (SNN) graph-based approach to find tripoles in a given spatio-temporal dataset.