Skip to content

MAINT Use memoryviews in _expected_mutual_info_fast #25575

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

MAINT Use memoryviews in _expected_mutual_info_fast #25575

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 12 additions & 13 deletions sklearn/metrics/cluster/_expected_mutual_info_fast.pyx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,20 +8,19 @@ import numpy as np
cimport numpy as cnp

cnp.import_array()
ctypedef cnp.float64_t DOUBLE


def expected_mutual_information(contingency, int n_samples):
"""Calculate the expected mutual information for two labelings."""
cdef int R, C
cdef DOUBLE N, gln_N, emi, term2, term3, gln
cdef cnp.ndarray[DOUBLE] gln_a, gln_b, gln_Na, gln_Nb, gln_nij, log_Nnij
cdef cnp.ndarray[DOUBLE] nijs, term1
cdef cnp.ndarray[DOUBLE] log_a, log_b
cdef cnp.ndarray[cnp.int32_t] a, b
cdef cnp.float64_t N, gln_N, emi, term2, term3, gln
cdef cnp.float64_t[:] gln_a, gln_b, gln_Na, gln_Nb, gln_nij, log_Nnij
cdef cnp.float64_t[:] nijs, term1
cdef cnp.float64_t[:] log_a, log_b
cdef cnp.int32_t[:] a, b
#cdef np.ndarray[int, ndim=2] start, end
R, C = contingency.shape
N = <DOUBLE>n_samples
N = <cnp.float64_t>n_samples
a = np.ravel(contingency.sum(axis=1).astype(np.int32, copy=False))
b = np.ravel(contingency.sum(axis=0).astype(np.int32, copy=False))

Expand All @@ -35,20 +34,20 @@ def expected_mutual_information(contingency, int n_samples):
nijs = np.arange(0, max(np.max(a), np.max(b)) + 1, dtype='float')
nijs[0] = 1 # Stops divide by zero warnings. As its not used, no issue.
# term1 is nij / N
term1 = nijs / N
term1 = nijs.base / N
# term2 is log((N*nij) / (a * b)) == log(N * nij) - log(a * b)
log_a = np.log(a)
log_b = np.log(b)
# term2 uses log(N * nij) = log(N) + log(nij)
log_Nnij = np.log(N) + np.log(nijs)
# term3 is large, and involved many factorials. Calculate these in log
# space to stop overflows.
gln_a = gammaln(a + 1)
gln_b = gammaln(b + 1)
gln_Na = gammaln(N - a + 1)
gln_Nb = gammaln(N - b + 1)
gln_a = gammaln(a.base + 1)
gln_b = gammaln(b.base + 1)
gln_Na = gammaln(N - a.base + 1)
gln_Nb = gammaln(N - b.base + 1)
gln_N = gammaln(N + 1)
gln_nij = gammaln(nijs + 1)
gln_nij = gammaln(nijs.base + 1)
# start and end values for nij terms for each summation.
start = np.array([[v - N + w for w in b] for v in a], dtype='int')
start = np.maximum(start, 1)
Expand Down