The crqa, R package, allows users to conduct a wide range of recurrence-based analyses on single (e.g., auto-recurrence) and multivariate time series (e.g., multidimensional cross-recurrence quantification), examine coupling properties underlying leader-follower relationships (i.e., diagonal-profile methods), as well as, track the evolution of recurrence rate over the time course (i.e., windowed methods).
# You can install the latest version of crqa on CRAN by running: install.packages("crqa") # Or for the development version from GitHub: # install.packages("devtools") ::install_github("morenococo/crqa")devtools
crqa comes with some data that can be used to test and study the different functions therein.
data(crqa) # load the data
First, specify the arguments that will be used in the crqa, core function.
## parameter setting = 1; embed = 1; rescale = 0; radius = 0.0001; delay = 0; mindiagline = 2; minvertline = 2; normalize = 1; whiteline = FALSE; recpt = FALSE; tw = "both"; method = 'rqa'; metric = 'euclidean'; side = "categorical"datatype
Then, run crqa on a nursery rhyme “The wheels on the bus” by Verna Hills: a vector of 120 strings (i.e., the words of the song),
= crqa(text, text, delay, embed, rescale, radius, normalize, ans mindiagline, minvertline, tw, whiteline, recpt, side, method, metric, datatype)
Have a look at the output, which contains different measures extracted from the recurrence plot (RP), and the RP itself, which can be plotted using the plotRP function.
Cross-recurrence extends univariate recurrence analysis into a bivariate analysis that allows quantification of the temporal coupling properties of two time series. We use eye-tracking data, 2,000 observations of six possible screen locations that are looked at by a dyad engaged into a joint task.
= eyemovement$listener listener = eyemovement$narrator narrator = 1; embed = 1; rescale = 0; radius = .01; delay = 0; mindiagline = 2; minvertline = 2; normalize = 0; whiteline = FALSE; recpt = FALSE; side = "both" tw = 'crqa'; metric = 'euclidean'; method = "categorical" datatype = crqa(narrator, listener, delay, embed, rescale, radius, normalize, ans mindiagline, minvertline, tw, whiteline, recpt, side, method, metric, datatype)
From cross-recurrence plots is possible to extract the diagonal cross-recurrence profiles (DCRPs) and use them to capture leader-follower-relationships.
= round( seq(-3300,3300,33)/1000, digit = 2) ## construct the time-course for the diagonal profile timecourse = drpfromts(narrator, listener, windowsize = 100, res radius = 0.001, delay = 1, embed = 1, rescale = 0, normalize = 0, mindiagline = 2, minvertline = 2, tw = 0, whiteline = F, recpt = F, side = 'both', method = 'crqa', metric = 'euclidean', datatype = 'continuous') ## visualise the diagonal-profile = res$profile*100 ## extract it from the res object profile plot(timecourse, profile, type = "l", lwd = 2.5, xlab = "Lag (seconds)", ylab = "Recurrence Rate %")
Multidimensional cross-recurrence quantification analysis allows for the computation of cross-recurrences between two multidimensional time-series. We use hand-movement data from a complex LEGO joint construction task. The dataframe comprises of 5,799 observations.
# reduce the dimensionality of the time series to make the computation faster # handset = handmovement[1:3000, ] = handmovement[1:1000, ] handset = cbind(handset$P1_TT_d, handset$P1_TT_n) P1 = cbind(handset$P2_TT_d, handset$P2_TT_n) P2 = 5; embed = 2; rescale = 0; radius = .1; delay = 0; mindiagline = 10; minvertline = 10; normalize = 0; whiteline = FALSE; recpt = FALSE; side = "both" tw = 'mdcrqa'; metric = 'euclidean'; method = "continuous" datatype = crqa(P1, P2, delay, embed, rescale, radius, normalize, ans mindiagline, minvertline, tw, whiteline, recpt, side, method, metric, datatype) = ans$RP RP = unlist(ans[1:10]) results print(results)