Comparison of Data Mining Techniques for Predicting Compressive Strength of Environmentally Friendly Concrete

Abstract

With its growing emphasis on sustainability, the construction industry is increasingly interested in environmentally friendly concrete produced using alternative and/or recycled waste materials. However, the wide application of such concrete is hindered by lack of understanding of the impacts of these materials on concrete properties. This research investigates and compares the performance of nine data mining models in predicting the compressive strength of a new type of concrete containing three alternative materials as fly ash, Haydite® lightweight aggregate, and Portland limestone cement. These models include three advanced predictive models (multilayer perceptron, support vector machines, and Gaussian processes regression), four regression tree models (M5P, REPTree, M5-Rules, and decision stump), and two ensemble methods (additive regression and bagging) with each of the seven individual models used as base classifier. The analytical results show that with appropriate parameter settings all of these models except for decision stump achieved acceptable prediction performance. The ensemble methods improved the prediction accuracy of the four regression tree models, but had less success on the other three advance predictive models. The individual Gaussian processes regression model as well as its related ensemble models reached the highest prediction accuracy in comparison groups. The results of this paper offer valuable insights to improving the use of these models for property prediction of concrete.

Mixture Optimization of Recycled Aggregate Concrete Using Hybrid Machine Learning Model

“…In contrast to single- or multi-variable probability distribution in which a scalar or a vector is mapped, a process describes the properties of functions [ 29 ]. Therefore, a GP is defined as a probability distribution of functions, P(f), where P(f) has a Gaussian distribution [ 30 ]. GPs are parametrized with mean and covariance by analogy with Gaussian distribution, whereas mean and covariance for GP are functions [ 31 ].…”

Estimating Compressive Strength of High Performance Concrete with Gaussian Process Regression Model

“…GPR presents a probabilistic, nonparametric supervised learning approach for generalizing nonlinear and complex function mapping hidden in data sets. This approach has recently received huge attention of researchers in various study disciplines [4,22]. GPR is very efficient to handle nonlinear data due to the use of kernel functions.…”

“…Furthermore, the compressive strength is often considered as the most important property of HPC; other concrete properties such as elastic modulus, water tightness, and impermeability appear to have direct relationships with compressive strength [3]. Hence, the compressive strength is commonly utilized as the main criterion in defining the required quality of concrete [4].…”

Multivariate regression models in estimating the behavior of FRP tube encased recycled aggregate concrete

“…Various analytical and modeling methods have been applied in predicting concrete properties, such as regression analysis [18], neural network [19], fuzzy logic [20], computer programming [21], and other data mining algorithms [22,23]. These multiple data analytical methods have been widely used to predict concrete mechanical properties, including compressive strength [24,25], structural capacity [26,27], as well as structural or shrinkage behaviour [28,29]. These methods adopted various IVs in predicting the target RRVs, such as the mix design involving environmentally friendly or "green" concrete materials [30,31].…”

“…Although the desired R 2 value depends on the decision-making context or research objectives and it could vary from 10% to 99% [40],a fairly highR 2 value is expected in the predictionof concrete properties. For example, it is not uncommon to see R 2 value higher than 0.9000 according to the study of Omran et al [24].So far these non-linear or mixed statistical methods have not been widely applied in evaluating concrete properties, especially in FRP-confined RAC specimens to improve the predication accuracy. Researchers believe that these statistical methods could further quantify the effects of multiple IVs in FRP-confined concrete properties.…”