15.2 Multiplex Recurrence Networks

Consider the three time series to be part of a multi-layer recurrence network. Common properties of the multiplex network are inter-layer mutual information and edge overlap can be calculated using function casnet::mrn(). One problem, the networks have to be all of the same size (same number of nodes, a multivariate time series), but here we have reconstructed the phase space using different embedding parameters… let’s choose one set of parameters for all time series.

emDim <- mean(c(p1$optimDim,p2$optimDim,p3$optimDim))
emLag <- median(c(p1$optimLag,p2$optimLag,p3$optimLag))

RNs <- plyr::llply(1:3, function(r) rn(y1 = series[,r], emDim = emDim, emLag = emLag, emRad = NA, targetValue = 0.05))
layers <- plyr::llply(RNs, function(r) igraph::graph_from_adjacency_matrix(r, mode="undirected", diag = FALSE))
names(layers) <- c("g1","g2","g3")

mrn(layers = layers)

A variety of plots can be created using casnet::mrn_plot()

# Simple
mrn_plot(layers = layers, showEdgeColourLegend =TRUE)
mrn_plot(layers = layers, MRNweightedBy = "EdgeOverlap", showEdgeColourLegend =TRUE)

# Include picture of Layers 
mrn_plot(layers = layers, RNnodes = TRUE)
mrn_plot(layers = layers, RNnodes = TRUE,MRNweightedBy = "EdgeOverlap", showEdgeColourLegend =TRUE)