Empirical comparison of correlation measures and pruning levels in complex networks representing the global climate system

Pelan, Alex; Steinhaeuser, Karsten; Chawla, Nitesh V.; Pitts, D. A. de Alwis; Ganguly, Auroop R.

doi:10.1109/cidm.2011.5949305

Cited by 7 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This implies that these studies have been limited to similarities captured by correlations among the variables under study. In this paper, we show that other distance functions can also be used for the construction of climate networks, an idea that was similarly proposed earlier by Pelan et al [34]. In the latter work, the authors used six different distance functions and analyzed the effect of edge density on the topology and the clustering predictive capability of the networks.…”

Section: Functional Climate Networksupporting

confidence: 55%

The effect of time series distance functions on functional climate networks

Ferreira

Macau

et al. 2021

Eur. Phys. J. Spec. Top.

View full text Add to dashboard Cite

Complex network theory provides an important tool for the analysis of complex systems such as the Earth's climate. In this context, functional climate networks can be constructed using a spatiotemporal climate dataset and a suitable time series distance function. The resulting coarse-grained view on climate variability consists of representing distinct areas on the globe (i.e., grid cells) by nodes and connecting pairs of nodes that present similar time series. One fundamental concern when constructing such a functional climate network is the definition of a metric that captures the mutual similarity between time series. Here we study systematically the effect of 29 time series distance functions on functional climate network construction based on global temperature data. We observe that the distance functions previously used in the literature commonly generate very similar networks while alternative ones result in rather distinct network structures and reveal different long-distance connection patterns. These patterns are highly important for the study of climate dynamics since they generally represent pathways for the long-distance transportation of energy and can be used to forecast climate variability on subseasonal to interannual or even decadal scales. Therefore, we propose the measures studied here as alternatives for the analysis of climate variability and to further exploit their complementary capability of capturing different aspects of the underlying dynamics that may help gaining a more holistic empirical understanding of the global climate system.

show abstract

Section: Functional Climate Networksupporting

confidence: 55%

The effect of time series distance functions on functional climate networks

Ferreira

Macau

et al. 2021

Eur. Phys. J. Spec. Top.

View full text Add to dashboard Cite

show abstract

“…For example, local neighborhood methods are clearly dominant in the YouTube network, a trait commonly observed in social networks [20], [21], [22], [23]. In the climate network, the Jaccard coefficient performs especially well, probably due to spatial autocorrelation [10]; that is, geographically proximate locations tend to have similar climate. In the disease network, each interaction type was best captured by a different method.…”

Section: Resultsmentioning

confidence: 99%

“…Every edge type can overlap with every other; a single node pair may have up to seven edges. Additional information about this network and the raw data can be found in [10].…”

Section: Climate Networkmentioning

confidence: 99%

Multi-relational Link Prediction in Heterogeneous Information Networks

Davis

Lichtenwalter

Chawla

2011

2011 International Conference on Advances in Social Networks Analysis and Mining

Self Cite

118

View full text Add to dashboard Cite

Abstract-Many important real-world systems, modeled naturally as complex networks, have heterogeneous interactions and complicated dependency structures. Link prediction in such networks must model the influences between heterogenous relationships and distinguish the formation mechanisms of each link type, a task which is beyond the simple topological features commonly used to score potential links. In this paper, we introduce a novel probabilistically weighted extension of the Adamic/Adar measure for heterogenous information networks, which we use to demonstrate the potential benefits of diverse evidence, particularly in cases where homogeneous relationships are very sparse. However, we also expose some fundamental flaws of traditional a priori link prediction. In accordance with previous research on homogeneous networks, we further demonstrate that a supervised approach to link prediction can enhance performance and is easily extended to the heterogeneous case. Finally, we present results on three diverse, real-world heterogeneous information networks and discuss the trends and tradeoffs of supervised and unsupervised link prediction in a multi-relational setting.

show abstract

“…In this paper, we show that other distance functions can be used for the construction of climate networks. A similar idea was proposed by Pelan et al (2011). The authors used six different distance functions.…”

Section: Climate Networkmentioning

confidence: 99%

The Effect of Time Series Distance Functions on Functional Climate Networks

Ferreira¹,

Ferreira²,

Macau³

et al. 2019

Preprint

View full text Add to dashboard Cite

Network theory has established itself as an important tool for the analysis of complex systems such as the climate. In this context, climate networks are constructed using a spatiotemporal climate dataset and a time series distance function. It consists of representing each spatial area by a node and connecting nodes that present similar time series. One fundamental concern when building climate network is the definition of a metric that captures similarity between time series. The majority of papers in the literature use Pearson correlation with or without lag. Here we study the influence of 29 time series distance functions on climate network construction using global temperature data. We observed that the distance functions used in the literature generate similar networks in general while alternative ones generate distinct networks and exhibit different long-distance connection patterns (teleconnections). These patterns are highly important for the study of climate dynamics since they generally represent long-distance transportation of energy and can be used to forecast climatological events. Therefore, we consider that the measures here studied represent an alternative for the analysis of climate systems due to their capability of capturing different underlying dynamics, what may provide a better understanding of global climate.

show abstract

Empirical comparison of correlation measures and pruning levels in complex networks representing the global climate system

Cited by 7 publications

References 19 publications

The effect of time series distance functions on functional climate networks

The effect of time series distance functions on functional climate networks

Multi-relational Link Prediction in Heterogeneous Information Networks

The Effect of Time Series Distance Functions on Functional Climate Networks

Contact Info

Product

Resources

About