Integrated experimental and ANN–TOPSIS framework for performance optimization of Melissa officinalis natural dyeing process on wool yarns

This study integrates experimental dyeing, artificial neural network (ANN) modeling, and TOPSIS-based optimization of Melissa officinalis aerial-part extract on wool yarns using metallic, bio-, and binary mordants. Color strength (K/S), fastness, and ultraviolet protection factor (UPF) were evaluated. Iron and copper mordants enhanced color strength and fastness (grades 4–5), with maximum K/S values of 7.99 and 5.12, respectively. The Fe/TA binary system exhibited the highest UV protection (UPF = 136.24), associated with metal–polyphenol–fiber complexation. Biomordants improved dyeing performance, particularly in binary combinations. ANN accurately predicted experimentally measured K/S values (R² ≈ 0.98), capturing nonlinear relationships between treatment conditions and color strength. TOPSIS systematically ranked treatments by simultaneously considering multiple colorimetric criteria. Integrating experimental characterization, ANN prediction, and TOPSIS decision analysis provides a data-driven framework for evaluating dyeing performance, identifying suitable mordanting conditions, and supporting further optimization of functional natural-dyed wool textiles.