The OrdinalPatterns documentation string promises that the user can provide a comparator function (defaults to lt = isless) to determine how two elements of a state are deemed to be "equal" (useful to prevent bias when there are tied values).
This function lt gets stored in the encoding::OrdinalPatternEncoding field of the OrdinalPatterns struct. However, the function isn't actually used in encode. We did use it before, but after we transitioned to the formal encode/decode interface, we forgot to pass the argument on to the underlying sortperm! call.
The fix is easy. Here's the source code:
function encode(encoding::OrdinalPatternEncoding{m}, χ::AbstractVector) where {m}
if m != length(χ)
throw(ArgumentError("Permutation order and length of input must match!"))
end
perm = sortperm!(encoding.perm, χ)
return permutation_to_integer(perm)
end
TODO:
[ ] Pass on lt to sortperm! when computing the permutation pattern.
We should add a simple test for this case, like for the permutation entropy of ones(1000); one would give 0 (the one with isless) while the other one would give maximum (with the random comparison).
The
OrdinalPatterns
documentation string promises that the user can provide a comparator function (defaults tolt = isless
) to determine how two elements of a state are deemed to be "equal" (useful to prevent bias when there are tied values).This function
lt
gets stored in theencoding::OrdinalPatternEncoding
field of theOrdinalPatterns
struct. However, the function isn't actually used inencode
. We did use it before, but after we transitioned to the formalencode
/decode
interface, we forgot to pass the argument on to the underlyingsortperm!
call.The fix is easy. Here's the source code:
TODO:
lt
tosortperm!
when computing the permutation pattern.