Prediction of mechanical properties of concrete incorporating quarry dust,silica fume, steel or polypropylene fibres using relevance vector machine

Abstract

This paper verifies the applicability of relevance vector machine (RVM) based regression to predict mechanical properties of quarry dust based concrete. Mechanical properties include compressive strength, split tensile strength and flexural strength. This paper presents the findings from experimental studies on concrete incorporating quarry dust, silica fume, steel fibres or polypropylene fibres. Relevance vector machine (RVM) is based on a probabilistic framework and introduces a-priori over the model weights governed by a set of hyper-parameters, associated with weights, whose most probable values are iteratively estimated from the data. Three RVM models have been developed using MATLAB software for training and prediction of mechanical properties. RVM models have been trained with about 70% of the total data sets and tested with about 30% of the total data sets. It is observed that the predicted values from the RVM models are in good agreement with those of the corresponding experimental values.