Plot_GC_all: Plot High-Dimensional Granger causality Networks

In Marga8/HDGCvar: Granger Causality Testing in High Dimensional Vector Autoregressive Models

Plot High-Dimensional Granger causality Networks

Description

Plot High-Dimensional Granger causality Networks

Usage

Plot_GC_all(
  Comb,
  Stat_type = "FS_cor",
  alpha = 0.01,
  multip_corr = list(F, "bonferroni", gamm = c(1e-04, 0.1, 0.001), fdr.apf = c(0.05,
    0.6), verb = F),
  ...,
  cluster = list(F, 10, "black", 0.51, 1, 0)
)

Arguments

Comb	output from: HDGC_VAR_all_I0, HDGC_VAR_multiple_pairs_I0, HDGC_VAR_all, HDGC_VAR_multiple_pairs, HDGC_HVAR_all or HDGC_HVAR_multiple_pairs
Stat_type	either FS_cor (default), Asymp or Asymp_Robust respectively for F-stat small sample correction, standard Chi square test, standard chi square test with heteroscedasticity correction
alpha	the desired probability of type one error, default is 0.01.
multip_corr	A list: first element is logical, if TRUE a multiple testing correction using p.adjust is used. The second element of the list define the p.adjust.method=c("holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none")). If the second element gets the name "APF_FDR" then apf_fdr is called which uses empirical Bayes is called and a third and fourth elements in the mutip_corr list are required: gamm=c(a,b,c) requires a min (a), max (b) and step length (c) values to be set for the threshold on the p_values, apf_fdr requires one or two values: either (NULL,value) or (value,NULL) if one wants to have specified amount of average power (fdr) no matter fdr (average power). If both (value,value) are given, the calculated threshold will find the closest combination to both apf and fdr desired. The last element of the list is logical: verbose=TRUE if one wants to know how much apf/fdr the testing has.
...	all parameters for the network plot: see example and graph_from_adjacency_matrix documentation.
cluster	A list: first element is logical, if TRUE a cluster plot using cluster_edge_betweenness is plotted. Other elements are respectively: vertex.size, vertex.label.color,vertex.label.cex, vertex.label.dist, edge.curved (see graph_from_adjacency_matrix for details).

Value

a graph_from_adjacency_matrix network

References

Hecq, A., Margaritella, L., Smeekes, S., "Inference in Non Stationary High Dimensional VARs" (2020, check the latest version at https://sites.google.com/view/luca-margaritella )

Hecq, A., Margaritella, L., Smeekes, S., "Granger Causality Testing in High-Dimensional VARs: a Post-Double-Selection Procedure." arXiv preprint arXiv:1902.10991 (2019).

Newman, Mark EJ, and Michelle Girvan. "Finding and evaluating community structure in networks." Physical review E 69.2 (2004): 026113.

Quatto, Piero, et al. "Brain networks construction using Bayes FDR and average power function." Statistical Methods in Medical Research 29.3 (2020): 866-878.

Examples

## Not run: Plot_GC_all(Comb, "FS_cor",alpha=0.01,multip_corr=list(F), directed=T, layout.circle

Marga8/HDGCvar documentation built on May 25, 2024, 11:12 a.m.