Machine Learning Algorithms - Second Edition

Adjusted Rand Index

The Adjusted Rand Index measures the similarity between the original class partitioning (Y) and the clustering. Bearing in mind the same notation adopted in the previous scores, we can define the following:

• a: The number of pairs of elements belonging to the same partition in the class set, C, and to the same partition in the clustering set, K
• b: The number of pairs of elements belonging to different partitions in the class set, C, and to different partitions in the clustering set, K

If the total number of samples in the dataset is n, the Rand Index is defined as follows:

The corrected for chance version is the Adjusted ...