Predictive Modelling of Marshall Stability using Multi Linear Regression: A Comprehensive Approach

Marshall Stability is a crucial parameter in the field of pavement engineering, serving as a key indicator of the strength and durability of asphalt mixtures. Accurate estimation of Marshall Stability is essential for ensuring the quality and performance of road pavements. This paper proposes the development of Multi Linear Regression (MLR) models as a predictive tool for estimating Marshall Stability based on various input parameters. A comprehensive dataset comprising a range of asphalt mixture compositions and Marshall Stability values is utilized for model training and validation. The input parameters considered in the MLR models include aggregate properties, binder characteristics, and mixture volumetrics. Through systematic feature selection and model refinement processes, robust MLR models are constructed to capture the complex relationships between input variables and Marshall Stability. The performance of the developed models is evaluated using statistical metrics such as coefficient of determination (R-squared), mean absolute error (MAE), and root mean square error (RMSE). The results demonstrate the efficacy of the proposed MLR approach in accurately estimating Marshall Stability, thus providing a valuable tool for engineers and researcher.