GNR

GNR implements a hybrid metaheuristic that enhances Nuclear Reaction Optimization by embedding a genetic uniform crossover mechanism into its fusion phase for gene selection in cancer classification.

Key Features:

• Two-stage gene selection: An initial F-score filtering step reduces dimensionality, followed by GNR-driven optimization to identify compact, informative gene subsets.
• Genetic-embedded fusion strategy: A genetic uniform crossover mechanism is embedded into the fusion phase of Nuclear Reaction Optimization to enhance local exploitation while preserving global search capabilities.
• Compact gene subset identification: The method achieves 100% classification accuracy with significantly smaller gene subsets across multiple microarray cancer datasets.

Scientific Applications:

• Cancer classification from gene expression profiles: GNR selects informative gene subsets from high-dimensional, low-sample-size microarray datasets to improve classification accuracy and reduce model complexity in precision oncology.
• Microarray data analysis: The method has been evaluated on six widely used microarray cancer datasets using Support Vector Machines as classifiers with Leave-One-Out Cross-Validation.

Methodology:

GNR employs a two-stage process: initial filtering of genes using F-score, followed by optimization using Genetic-Embedded Nuclear Reaction Optimization. Support Vector Machines serve as the classifier, with performance assessed via Leave-One-Out Cross-Validation.