Pairwise Profile Comparison

When doing Ranking, arbitrary scores are hard to measure especially when we are using multiple annotators, then its better to use pairwise ranking algorithms to rank objects.

The issue proposes pairwise ranking for comparisons involving multiple annotators, but names no files, entry points, tests, or acceptance criteria. First clarify the intended ranking algorithm, inputs, outputs, and project scope; done would require an agreed design and a validated implementation.

Written by the indexing model from the issue text.