Modelling of Tensile Strength Ratio of Bituminous Concrete Mixes Using Support Vector Machines and M5 Model Tree

Bituminous pavements are greatly vulnerable to moisture damage. Most of the studies all over the world conclude that presence of moisture on road surface contributes significantly to pavement failures. Sustainable bituminous pavement needs high quality bitumen material. In roadway design, the tensile strength of bituminous pavements is an important engineering property. Indirect tensile strength (ITS) tests are performed on bituminous mixes for evaluating tensile strength ratio (TSR). Knowledge of TSR is helpful in determining the moisture susceptibility of bituminous concrete (BC) mixes for highway pavements. In the past different modelling techniques were developed to predict the TSR of bituminous mixes. In this study, support vector machines (SVM), M5 model tree and regression-based technique were used to estimate the TSR value of BC mixes. Two kernel functions, i.e. polynomial and radial-based kernel function were used with SVM model. Cross validation technique with 10-fold analysis was used for testing the models. SVM model with radial basis function (RBF) outclasses M5P model all in terms of correlation coefficient and R-2 values. This paper proposes a simple linear equation for the prediction of the TSR, which is based on ITS tests performed on different types of bituminous mixes.