/AdjustedRandCI

Compute Adjusted Rand Index and Confidence Interval. The confidence interval is computed using central limit theorem and the exact variance of the Adjusted Rand Index.

Primary LanguagePythonMIT LicenseMIT

AdjustedRandCI

Compute Adjusted Rand Index and Confidence Interval. The confidence interval is an approximate normal confidence interval using the exact variance of the Adjusted Rand Index.

For details see: Steinley, Douglas, Michael J. Brusco, and Lawrence Hubert. "The variance of the adjusted Rand index." Psychological methods 21.2 (2016): 261.

Dependencies

numpy,scipy,sklearn

Example

import numpy as np
from ari_ci import ari_ci

clusters = np.array([0,1,1,1,1,1,1,1,1,0,0,0,0])
trueclusters = np.array([0,1,1,1,1,0,0,1,1,1,0,0,0])

# print ARI, variance of ARI, lower CI, upper CI
print(ari_ci(clusters,trueclusters,alpha = 0.05))