Examples
Examples
Try the starting example
- Records: unique sample ID, partition A label, partition B label
- R1,A,X R2,A,X R3,B,Y R4,B,Y
- Adjusted Rand index
- 1
- Unadjusted Rand agreement
- 1
- Total distinct sample pairs
- 6
- Same-cluster pairs in both partitions
- 2
- Same-cluster pair counts in A and B
- 2, 2
Try another scenario
- Records: unique sample ID, partition A label, partition B label
- R1,A,X R2,A,Y R3,B,Y R4,B,Y
- Adjusted Rand index
- 0
- Unadjusted Rand agreement
- 0.5
- Total distinct sample pairs
- 6
- Same-cluster pairs in both partitions
- 1
- Same-cluster pair counts in A and B
- 2, 3
Calculation method
Calculation method
Adjusted Rand index
Exact integer chance-adjusted pair counts; degenerate identical pair relations return 1.
Unadjusted Rand agreement
Fraction of sample pairs whose same/different relation agrees.
Total distinct sample pairs
n(n−1)/2.
Same-cluster pairs in both partitions
Sum choose(cell count,2) over the partition contingency table.
Same-cluster pair counts in A and B
Sum choose(row count,2) and sum choose(column count,2).
Symbols and units
- a
- Records: unique sample ID, partition A label, partition B label
Use the units shown. A list keeps the order in which you enter its values. The symbol ^ means “raised to the power of”.
Assumptions & limits
Assumptions & limits
- Compares two supplied hard partitions of exactly the same records, without fitting clusters or assessing their real-world meaning. Label names are arbitrary; degenerate identical pair relations, including one record, use agreement 1.
Are these inputs private?
These calculations run in your browser. A shared link contains the values you choose to share. Only questions submitted to the separate AI finder are sent to OpenAI.
Further reading
scikit-learn: adjusted_rand_score ↗
Background reading; this does not imply endorsement or independent review of SumFizz.