Answers are grounded in this paper and cite their sources.
Understanding and predicting the mechanical behavior of woven denim fabrics is essential for optimizing design and production. This study presents a comparative and interpretable machine learning framework to predict tensile and tear strength in both warp and weft directions using intrinsic fabric structural and material parameters. A comprehensive industrial dataset, including yarn composition, fabric construction, and physical characteristics, was analyzed using Random Forest (RF) and Linear Regression (LR) models. The RF model consistently outperformed LR, achieving R2 values of 0.884 for warp tensile strength, 0.901 for warp tear strength, and 0.892 for weft tear strength, demonstrating its capacity to capture nonlinear interactions. Feature importance analysis revealed reed number as the most influential factor (RF score = 100), followed by elastane (42.19) and cotton content (42.15), emphasizing the dominant role of yarn density and fiber composition. Unlike prior studies that often focus on single properties or isolated features, this research integrates multi-directional mechanical properties with a heterogeneous industrial dataset, uncovering cross-directional dependencies and providing mechanistic insights. The findings contribute to the literature by establishing a robust, data-driven methodology that enables accurate prediction, interpretable analysis, and practical guidance for denim fabric optimization, bridging the gap between experimental testing and industrial textile design.
No takes yet. Share an insight, caveat, or question.
Polat et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: