scholarly journals Node Centrality Measures are a poor substitute for Causal Inference

2018 ◽  
Author(s):  
Fabian Dablander ◽  
Max Hinne
Keyword(s):  
Network Models ◽  
Directed Acyclic Graphs ◽  
Centrality Measures ◽  
Eigenvector Centrality ◽  
Diverse Range ◽  
Causal Influence ◽  
Node Centrality ◽  
Making Sense ◽  
Acyclic Graphs ◽  
Causal Framework

Network models have become a valuable tool in making sense of a diverse range of social, biological, and information systems. These models marry graph and probability theory to visualize, understand, and interpret variables and their relations as nodes and edges in a graph. Many applications of network models rely on undirected graphs in which the absence of an edge between two nodes encodes conditional independence between the corresponding variables. To gauge the importance of nodes in such a network, various node centrality measures have become widely used, especially in psychology and neuroscience. It is intuitive to interpret nodes with high centrality measures as being important in a causal sense. Here, using the causal framework based on directed acyclic graphs (DAGs), we show that the relation between causal influence and node centrality measures is not straightforward. In particular, the correlation between causal influence and several node centrality measures is weak, except for eigenvector centrality. Our results provide a cautionary tale: if the underlying real-world system can be modeled as a DAG, but researchers interpret nodes with high centrality as causally important, then this may result in sub-optimal interventions.

scholarly journals Adjustment for time-invariant and time-varying confounders in ‘unexplained residuals’ models for longitudinal data within a causal framework and associated challenges

2018 ◽  
Vol 28 (5) ◽  
pp. 1347-1364 ◽  
Author(s):  
KF Arnold ◽  
GTH Ellison ◽  
SC Gadd ◽  
J Textor ◽  
PWG Tennant ◽  
...  
Keyword(s):  
Directed Acyclic Graphs ◽  
Regression Methods ◽  
Modelling Framework ◽  
Acyclic Graphs ◽  
Confounding Variables ◽  
Confounder Adjustment ◽  
Causal Framework ◽  
Distal Outcome ◽  
Time Invariant ◽  
Insight Into

‘Unexplained residuals’ models have been used within lifecourse epidemiology to model an exposure measured longitudinally at several time points in relation to a distal outcome. It has been claimed that these models have several advantages, including: the ability to estimate multiple total causal effects in a single model, and additional insight into the effect on the outcome of greater-than-expected increases in the exposure compared to traditional regression methods. We evaluate these properties and prove mathematically how adjustment for confounding variables must be made within this modelling framework. Importantly, we explicitly place unexplained residual models in a causal framework using directed acyclic graphs. This allows for theoretical justification of appropriate confounder adjustment and provides a framework for extending our results to more complex scenarios than those examined in this paper. We also discuss several interpretational issues relating to unexplained residual models within a causal framework. We argue that unexplained residual models offer no additional insights compared to traditional regression methods, and, in fact, are more challenging to implement; moreover, they artificially reduce estimated standard errors. Consequently, we conclude that unexplained residual models, if used, must be implemented with great care.

scholarly journals Building-network: Concept, generation method and centrality analysis

2019 ◽  
Vol 2 ◽  
pp. 1-8
Author(s):  
Xiao Wang ◽  
Dirk Burghardt
Keyword(s):  
Network Theory ◽  
Network Models ◽  
Pattern Detection ◽  
Centrality Measures ◽  
Concept Generation ◽  
Eigenvector Centrality ◽  
Rural Regions ◽  
Network Concept ◽  
Centrality Analysis ◽  
Refinement Strategy

<p><strong>Abstract.</strong> Buildings are among the most important features of cities. In the suburban or rural regions, buildings are normally constructed along the roads, which forms the smooth and consistent patterns so that the building arrangements also can be described with network models. In previous studies, network theory has achieved good performance in cartography and GIS. In this paper, a study of a building-network is proposed, including the concepts, generation methods and centrality analysis. Firstly, with the constraint Delaunay triangulation and the refinement strategy by facing ratio, the building-network is generated by considering the buildings and the proximal segments as the nodes and segments of the network, respectively. Then, centrality analysis is applied on the building-network, aiming to reveal the crucial relationships among buildings, which is useful for understanding the structural properties of the complex network. Four different centrality measures, i.e. degree, closeness, betweenness, and eigenvector centrality, are calculated based on the building-networks. The buildings show different distribution effects and patterns under the four centrality measures. From the results, the degree centrality reveals the local centre of the region; closeness and eigenvector centrality have the ability to cluster buildings into different groups; while betweenness centrality can detect the linear patterns. Therefore, using network theory to analyse buildings can reveal some inner relationships of buildings and has great potential in the application of building pattern detection, classification, clustering and further generalization.</p>

scholarly journals Longitudinal symptomatic interactions in long-standing schizophrenia: a novel five-point analysis based on directed acyclic graphs

2021 ◽  
pp. 1-8
Author(s):  
Giusi Moffa ◽  
Jack Kuipers ◽  
Giuseppe Carrà ◽  
Cristina Crocamo ◽  
Elizabeth Kuipers ◽  
...  
Keyword(s):  
Network Models ◽  
Chronic Schizophrenia ◽  
Directed Acyclic Graphs ◽  
Psychotic Symptoms ◽  
Outpatient Services ◽  
Cross Sectional ◽  
Affective Symptoms ◽  
Acyclic Graphs ◽  
Affective Disturbance ◽  
The Uk

Abstract Background Recent network models propose that mutual interaction between symptoms has an important bearing on the onset of schizophrenic disorder. In particular, cross-sectional studies suggest that affective symptoms may influence the emergence of psychotic symptoms. However, longitudinal analysis offers a more compelling test for causation: the European Schizophrenia Cohort (EuroSC) provides data suitable for this purpose. We predicted that the persistence of psychotic symptoms would be driven by the continuing presence of affective disturbance. Methods EuroSC included 1208 patients randomly sampled from outpatient services in France, Germany and the UK. Initial measures of psychotic and affective symptoms were repeated four times at 6-month intervals, thereby furnishing five time-points. To examine interactions between symptoms both within and between time-slices, we adopted a novel technique for modelling longitudinal data in psychiatry. This was a form of Bayesian network analysis that involved learning dynamic directed acyclic graphs (DAGs). Results Our DAG analysis suggests that the main drivers of symptoms in this long-term sample were delusions and paranoid thinking. These led to affective disturbance, not vice versa as we initially predicted. The enduring relationship between symptoms was unaffected by whether patients were receiving first- or second-generation antipsychotic medication. Conclusions In this cohort of people with chronic schizophrenia treated with medication, symptoms were essentially stable over long periods. However, affective symptoms appeared driven by the persistence of delusions and persecutory thinking, a finding not previously reported. Although our findings as ever remain hostage to unmeasured confounders, these enduring psychotic symptoms might nevertheless be appropriate candidates for directly targeted psychological interventions.

Centrality Measures

2018 ◽  
Author(s):  
Ginestra Bianconi
Keyword(s):  
State Of The Art ◽  
Centrality Measures ◽  
Eigenvector Centrality ◽  
Multilayer Networks ◽  
Comprehensive Description ◽  
Multilayer Network ◽  
Suitable Definition ◽  
Katz Centrality ◽  
Pagerank Centrality

Defining the centrality of nodes and layers in multilayer networks is of fundamental importance for a variety of applications from sociology to biology and finance. This chapter presents the state-of-the-art centrality measures able to characterize the centrality of nodes, the influences of layers or the centrality of replica nodes in multilayer and multiplex networks. These centrality measures include modifications of the eigenvector centrality, Katz centrality, PageRank centrality and Communicability to the multilayer network scenario. The chapter provides a comprehensive description of the research of the field and discusses the main advantages and limitations of the different definitions, allowing the readers that wish to apply these techniques to choose the most suitable definition for his or her case study.

scholarly journals Topological measures for identifying and predicting the spread of complex contagions

2021 ◽  
Vol 12 (1) ◽  
Author(s):  
Douglas Guilbeault ◽  
Damon Centola
Keyword(s):  
Social Networks ◽  
Shortest Path ◽  
Empirical Data ◽  
Path Length ◽  
Centrality Measures ◽  
Social Phenomena ◽  
Node Centrality ◽  
Network Connectedness ◽  
Topological Measures ◽  
Definition Of

AbstractThe standard measure of distance in social networks – average shortest path length – assumes a model of “simple” contagion, in which people only need exposure to influence from one peer to adopt the contagion. However, many social phenomena are “complex” contagions, for which people need exposure to multiple peers before they adopt. Here, we show that the classical measure of path length fails to define network connectedness and node centrality for complex contagions. Centrality measures and seeding strategies based on the classical definition of path length frequently misidentify the network features that are most effective for spreading complex contagions. To address these issues, we derive measures of complex path length and complex centrality, which significantly improve the capacity to identify the network structures and central individuals best suited for spreading complex contagions. We validate our theory using empirical data on the spread of a microfinance program in 43 rural Indian villages.

scholarly journals Directed acyclic graphs: An under-utilized tool for child maltreatment research

2019 ◽  
Vol 91 ◽  
pp. 78-87 ◽  
Author(s):  
Anna E. Austin ◽  
Tania A. Desrosiers ◽  
Meghan E. Shanahan
Keyword(s):  
Child Maltreatment ◽  
Directed Acyclic Graphs ◽  
Acyclic Graphs
Sign in / Sign up

Export Citation Format

Share Document