Convenience function for evaluating edge detection performance of CV.Vote model fits.

cvPerf(cvOut, truth, method, givenX = FALSE, Xindex = NULL, Yindex = NULL,
  aszero = 1e-06)

Arguments

cvOut

Output from cvVote.

truth

A list of numeric matrices named 'xy' and 'yy' encoding the true model parameters (or an adjacency matrix).

method

Character vector indicates cvOut came from spaceMap or SPACE model. See parameter method in cvVote.

givenX

Logical (defaults to FALSE). Set to TRUE for method == "space" when cvOut had non-null X input to cvVote.

Xindex

Integer vector of indices of X variables in partial correlation matrix. Defaults to NULL, but must be non-null when method == "space" & givenX = TRUE.

Yindex

Integer vector of indices of Y variables in partial correlation matrix. Defaults to NULL, but must be non-null when method == "space" & givenX = TRUE.

aszero

A numeric value specifying the point at which a parameter estimate should be effectively considered of zero value.

See also

reportJointPerf, reportPerf