partial cross mapping

Usage

# S4 method for class 'data.frame'
pcm(
  data,
  cause,
  effect,
  conds,
  libsizes,
  E = 3,
  tau = 0,
  k = E + 1,
  theta = 1,
  algorithm = "simplex",
  lib = NULL,
  pred = NULL,
  threads = length(libsizes),
  parallel.level = "low",
  bidirectional = TRUE,
  cumulate = FALSE,
  progressbar = TRUE
)

Arguments

data: observation data.
cause: name of causal variable.
effect: name of effect variable.
conds: name of conditioning variables.
libsizes: number of time points used in prediction.
E: (optional) embedding dimensions.
tau: (optional) step of time lags.
k: (optional) number of nearest neighbors used in prediction.
theta: (optional) weighting parameter for distances, useful when algorithm is smap.
algorithm: (optional) prediction algorithm.
lib: (optional) libraries indices.
pred: (optional) predictions indices.
threads: (optional) number of threads to use.
parallel.level: (optional) level of parallelism, low or high.
bidirectional: (optional) whether to examine bidirectional causality.
cumulate: (optional) serial or cumulative computation of partial cross mapping.
progressbar: (optional) whether to show the progress bar.

Value

A list

pxmap: partial cross mapping results
xmap: cross mapping results
varname: names of causal and effect variable
bidirectional: whether to examine bidirectional causality

References

Leng, S., Ma, H., Kurths, J. et al. Partial cross mapping eliminates indirect causal influences. Nat Commun 11, 2632 (2020).

Examples

sim = logistic_map(x = 0.4,y = 0.4,z = 0.4,step = 45,
                   beta_xy = 0.5, beta_xz = 0,
                   beta_yx = 0, beta_yz = 0.5,
                   beta_zx = 0, beta_zy = 0)
pcm(sim,"x","z","y",libsizes = seq(5,35,5),E = 10,k = 7,threads = 1)
#> 
Computing: [========================================] 100% (done)                         
#> 
Computing: [========================================] 100% (done)                         
#> -------------------------------------- 
#> ***partial cross mapping prediction*** 
#> -------------------------------------- 
#>   libsizes      x->z      z->x
#> 1        5 0.2549662 0.1687161
#> 2       10 0.1025583 0.1287902
#> 3       15 0.2345731 0.1407319
#> 4       20 0.1310661 0.2028892
#> 5       25 0.1266602 0.2397178
#> 6       30 0.1147702 0.2771978
#> 7       35 0.1245560 0.5516479
#> 
#> ------------------------------ 
#> ***cross mapping prediction*** 
#> ------------------------------ 
#>   libsizes      x->z      z->x
#> 1        5 0.8099271 0.6423512
#> 2       10 0.8818607 0.6830390
#> 3       15 0.8961755 0.6924374
#> 4       20 0.8961413 0.6965852
#> 5       25 0.8948321 0.7039838
#> 6       30 0.8941432 0.7064543
#> 7       35 0.9346800 0.7623188