The Temporal Logic of Causal Structures
[Back to full results]
Significance testing
Our tests as well as the Granger tests each return a set of causal hypotheses with some associated test statistics. One must still choose a threshold at which a result will be called significant (or causal). Here we include the set of histograms for each pattern depicting the number of results with a given test value. For the granger test, which returned both p-values and f-statistics, we graph both.
Links to graphs below
Our tests:
One (scatter gather), Two (two separate scatter gather graphs), Three (a chain of scatter gather relationships), Four (a binary tree), and Five (a long chain of neurons).
Granger tests:
One, Two, Three, Four, and Five.