Features

Model Languages

Binary (bin) – Features contribute as 0 or 1 (presence/absence)
Ternary (ter) – Features contribute as -1, 0, or +1 (direction matters)
Ratio – Score is computed as a ratio of two feature groups

Search Algorithms

Genetic Algorithm (GA) – Population-based evolutionary search with crossover, mutation, and selection
Beam Search – Deterministic heuristic that grows signatures incrementally
Data-driven (Da) – Analytical approach for optimal feature selection

Jury / Ensemble Voting

Build an ensemble of expert models from the best individuals across generations
Majority voting – Weighted expert consensus with configurable threshold
Consensus voting – Requires minimum agreement level to predict
Rejection – Samples below confidence threshold are abstained (class 2)
Per-sample vote matrix visualization and concordance analysis

Evaluation

K-fold cross-validation with AUC, accuracy, sensitivity, specificity
Train/test split with independent evaluation
Confusion matrices with rejection class support
Feature importance ranking via permutation

Web Application

Project and dataset management with user authentication
Parameter configuration with admin-level defaults
Job queue with real-time progress monitoring
Interactive results: heatmaps, violin plots, generation tracking, population explorer
Jury visualization: concordance charts, sample predictions, vote matrix heatmap
Project sharing between users (viewer/editor roles)