Quantitative Evaluation and Machine Learning‐Based Prediction of Linear Density and Tensile Strength in Glass Fiber‐Reinforced Tapes for Enhanced Quality Control

ABSTRACT Glass fiber‐reinforced tapes are increasingly used in high‐performance sectors due to their high specific strength and dimensional stability. Ensuring quality requires accurate assessment of tensile strength (TS) and linear density (Tex), which are influenced by microstructural parameters such as fiber weight fraction (FWF) and fiber diameter (FD). Traditional experimental and linear statistical models, however, struggle to capture the complex nonlinear relationships governing these properties. To address this, the present study integrates experimental evaluation with machine learning (ML)‐based prediction of TS and Tex. A dataset of 100 samples was generated and analyzed using eight supervised ML algorithms. Results demonstrated that nonlinear ML models substantially outperformed linear approaches in predictive accuracy. Support vector regression (SVR) achieved the highest performance for both TS ( R 2 = 0.995, MAE = 1.524 MPa, RMSE = 1.892 MPa) and Tex ( R 2 = 0.979, MAE = 802.559, RMSE = 1024.195). Multilayer Perceptron (MLP) followed closely, while ensemble models, Random Forest Regressor (RFR) and XGBoost Regressor (XGBR) provided strong accuracy with moderate error. Feature attribution analyses revealed FWF as the dominant predictor, with FD exerting secondary influence. This study highlights the advantages of nonlinear ML approaches, establishing a predictive framework for quality control in industrial glass fiber‐reinforced tapes.