Description Usage Arguments Value
CCM predictability for a given lag is only considered if the analysis is convergent. Convergence is determined by Wilcoxon rank-sum test on the smallest and largest library sizes. If CCM skill for the largest library sizes is higher (p < 0.01) than for the lowest library sizes, then the analysis is potentially convergent. Additionally, the median CCM curve over increasing library sizes must be satisfactorily fitted by a convergent exponential function. If both these criteria are met, then the analysis is labelled convergent. The only grouping variable in this summary function is the lag, so all further groupings must be handled outside this function.
1 | directionalcausaltest(res, library.size = max(res$library.size))
|
res |
A data frame containing the result of a tstools::ccm_lagged() call. |
library.size |
The library size to summarise over. Defaults to the largest available library size. |
The difference between summed negative median CCM skills and summed positive median CCM skills. If the sum > 0, then the analysis passes the lag test.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.