Research pilot applications are open. Apply for pilot →
Evaluation & synthesis

Comparing analytical runs

Compare algorithms, parameterizations, and analytical spaces through shared views and metrics.

What can be compared

Ounias supports side-by-side comparison of up to four runs, including runs created with different algorithms or analytical configurations.

Dataset-to-dataset comparison is a separate planned capability.

Comparison views

  • Review method-specific metrics and standardized evaluation pillars.
  • Use radar and bar plots to inspect contrasts between runs.
  • Compare alternatives in feature space and embedded space.
  • Return to individual runs for supporting plots, explanations, and artifacts.

Interpreting differences

Comparing scores across different methods can reveal strengths and tradeoffs, but it does not imply that every metric has identical meaning or that a higher value represents the most relevant research answer.