ScoreQuant documentation

ScoreQuant chooses hard bins from model scores to retain Fisher information at a reference parameter point. Use optimize_partition for labels on a fixed sample, or fit_quantizer for a reusable rule. Estimated scores produce a surrogate information measure.

Start here

• Install and run — a complete first workflow.
• Choose a workflow — task, source, provider, criterion and solver.
• API guide and symbol reference — inputs, outputs and errors.

Understand and reproduce

• Walkthroughs — four applied problems, with models and interactive experiments.
• Runnable examples — scripts and committed results.
• The book — derivations, assumptions and limits.
• Related work, bibliography and glossary.

Detailed evidence: FlowCyt, HEP, and the synthetic gallery.