Analysis and prediction of plastic waste composite construction material properties using machine learning techniques

Abstract

AbstractThe use of plastic‐based composites as building materials may be one method of recycling plastic waste. Plastic pollution has emerged as a global issue since our planet is becoming increasingly engulfed in plastic waste. In order to preserve a long‐term sustainable ecosystem, alternative methods of recycling this plastic trash would be extremely beneficial. In this study, the six attributes, which are ingredients used to make plastic composite construction material, are used to predict five target variables or properties of plastic composite construction material, such as compressive strength, flexural strength, split tensile strength, density, and water absorption. Five machine learning techniques random forest, Decision Tree, Ridge, Lasso, and Linear regression have been used to get the best prediction model for the properties of plastic composite construction material. The correlation method has been used to analyze the relationships among the properties and ingredients of plastic composite construction material. Analytical results for K‐fold cross‐validation show that random forest and decision tree form the best predictive models, having greater prediction accuracy than the rest of the machine learning algorithms. This pioneering work provides a simple and convenient method for predicting various properties of plastic‐based composite construction material and finding the impact of each composition mix on various properties.