Confusion Matrix vs Correlation Table as Visualization

[BradKML]

Is your feature request related to a problem? Please describe.
Currently, the partition comparisons are in a sense "just an index"
https://cdlib.readthedocs.io/en/latest/reference/evaluation.html#partition-comparisons

It might be a good idea to visualize this with either Confusion Matrix (if the two partitions have the same quantity of clusters) OR Correlation Table (if the two partitions have different quantity of clusters)

Reference.

• http://www.sthda.com/english/wiki/visualize-correlation-matrix-using-correlogram

Describe the solution you'd like

1. For a compact correlation table, it should display non-overlapping partitions A and B as an MxN grid rather than (M+N)x(M+N).
2. Since the data is binary rather than continuous, there should be a metric for assessing correlatedness. Callback on Documentation on existing tables of External Validation methods #147
3. For overlapping partitions C and D, a standard Correlation/Confusion Matrix would be useful to see how intra-partition verlap works.