Module aggregation 

Fitness aggregation models for interactive evaluation

This module provides various statistical models for converting user feedback (ratings, comparisons, selections) into fitness values suitable for evolution.

Available Models

• DirectRating: Simple average of user ratings
• Elo: Classic Elo rating system from pairwise comparisons
• BradleyTerry: Maximum likelihood estimation for pairwise data
• ImplicitRanking: Bonus/penalty system from batch selections

Uncertainty Quantification

All models support uncertainty estimation via get_fitness_estimate(), which returns a FitnessEstimate with variance and confidence intervals.

Structs

CandidateStats

Statistics tracked for each candidate

ComparisonRecord

Record of a pairwise comparison

FitnessAggregator

Aggregates partial/incremental feedback into fitness estimates

Enums

AggregationModel

Aggregation model for converting user feedback to fitness