Talk:Convergent cross mapping
 | This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | ||||||||||
| |||||||||||
Update for CCM
[edit]Dear Wikiverse,
After review of the CCM page, based on recent developments, and, to clarify and improve the citations, I am proposoing to update the page.
For instance, citation of an opinion article from a non-peer reviewed magazine does not meet Wikipedia standards for peer-reviewed scientific literature as the basis of objective article references. The citation of Cenys is not quite appropriate as this paper does not relate to convergent cross mapping, rather to estimation of interaction coefficients.
I also propose to rewrite the Theory section. For instance, the use of "shadow" manifold is ambiguous. In fact, there is no requirement for a "shadow" manifold, the algorithm works equally well on an observed multivariate state-space.
The algorithm section can be improved. "Creating weights" is an internal aspect of the simplex projection for which there is no parametric control, thereby it is not really an actual "step" the user undertakes. A crucial and often ignored aspect of CCM is the random sampling (which is the aspect that qualfies CCM as "statstical test") that averages the result of state-space projected cross mappings. In fact, the last step listed in the algorithm is not quite right, as this averaging is done over the Pearson correlations of each ensemble. Also, significance testing is an important aspect for practical use of the metric.
There have also been significant extensions to CCM in the last few years. These should be cited.
Thank you!
J Park (talk) 03:47, 14 September 2022 (UTC)
Reintroducing references about fundamental previous work
[edit]Dear community,
in an earlier version I had added some information on the history of this approach. This information has been removed in the meantime. Like I reason below, I propose to introduce it again.
Why? What is described in the section "Theory" was known already before the publication of the convergent cross mapping statistics in 2012[1]. It was first described by Cenys et al. in 1991[2]. It was furthermore used in a series of statistical approaches which were published before 2012 (see for example,[3][4][5]). I think this information should be incorporated in the section "Theory". This would also help to support this text with further references from peer-reviewed publications.
I agree with @J Park that the Cenys 1991 paper does not relate to the special technique of convergent cross mapping. However, it was the first to describe the criterion which is outlined in the section "Theory" and is used for the convergent cross mapping algorithm. Moreover, the convergent cross mapping algorithm was not the first to use this criterion.
- ^ Sugihara, George; May, Robert; Ye, Hao; Hsieh, Chih-hao; Deyle, Ethan; Fogarty, Michael; Munch, Stephan (2012). "Detecting Causality in Complex Ecosystems". Science. 338 (6106): 496–500. Bibcode:2012Sci...338..496S. doi:10.1126/science.1227079. PMID 22997134. S2CID 19749064.
- ^ A. Cenys, G. Lasiene, and K. Pyragas, Physica D 52, 332, 1991
- ^ S. J. Schiff, P. So, T. Chang, R. E. Burke, and T. Sauer, Phys. Rev. E 54, 6708, 1996
- ^ J. Arnhold, P. Grassberger, K. Lehnertz, and C. E. Elger, Physica D 134, 419, 1999
- ^ D. Chicharro and R. G. Andrzejak, Phys. Rev. E 80, 026217, 2009