Using the output of compute_ccm(), determine which
links are significant. Significant links (x "causes" y) are where:
ccm rho for the actual time series x and y is greater than the
null_quantile level of the surrogate data (at the largest library size)
the increase in rho for the actual time series x and y between the
smallest and largest library sizes is greater than delta_rho_threshold
compute_ccm_links(ccm_results, null_quantile = 0.975, delta_rho_threshold = 0.1)
| ccm_results | the output from |
|---|---|
| null_quantile | the quantile of the surrogate which we desire the actual
ccm rho to exceed (the default value of |
| delta_rho_threshold | the absolute increase which we desire the actual ccm rho values to exceed, when comparing the smallest and largest library sizes |
A tibble with the filtered significant links, along with the values
of the test statistics that were computed (delta_rho and
rho_minus_upper_q_null)