Open Phoebe715 opened 8 months ago
What should be the explanation if r is negative
What should be the explanation if r is negative
the r being reported is the pearson correlation coefficient. negative correlations imply that the variables are inversely related (i.e., as x increases, y decreases.)
You can change the tau argument - ccm(X, Y, tau, E, L) you can write a for loop for example,
for tau in [1, 2, 3]:
ccm_instance = ccm(X, Y, tau, E, L)
# rest of your code
As mentioned by scibanez, negative r means inversely related. in the context of CCM, this just means there is no causal relationship between the two. You want an r
that is positive and significant
@Phoebe715 To add to my previous explanation on negative r, in the supplemental materials of Sugihara et al's paper, they explained that negative r are truncated to 0. I'm planning to include this as a feature - truncation of the correlation value to only be between 0 and 1
If I need to study multiple lag times how should I accomplish this? For example tau=2, 3 or 100.