Green concrete with oil palm shell aggregate: usage of the chaos game-based tree algorithm

In tropical areas, the waste product known as oil palm shell is often generated throughout the palm oil refining process. It is costly and time-consuming to use OPS\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$OPS$$\end{document} to measure the compressive strength of concrete during manufacturing. The goal of the current study is to offer novel hybrid models that integrate the ideas of the Random Forests regression model with optimization techniques, such as the Chaos game algorithm and Artificial rabbit algorithm. The purpose of these models is to forecast the oil palm shell lightweight aggregate concrete's uniaxial compressive strength precisely. The results show that both models operate well when it comes to UCS\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$UCS$$\end{document} predicting. The integration of random forest and artificial rabbit algorithm and its combination with chaos game algorithm has a coefficient of determination value of 0.9836 during training and 0.9818 during testing. The random forest-artificial rabbit algorithm has a R2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}<^>{2}$$\end{document} value of 0.9705 during training and 0.9646 during testing. In the training and testing stages, random forest-chaos game algorithm values were more suitable than the other model's values for root mean square error and mean absolute error analysis. The root mean square error decreased from 1.6949 in the random forest-artificial rabbit algorithm to 1.3193 in the random forest-chaos game algorithm during training and from 1.5621 to 1.2056 during testing. As a comprehensive index, objective index random forest-chaos game algorithm is 1.1085, whereas, for random forest-artificial rabbit algorithm, it is 1.4531, indicating greater performance, where can be considered as a tool in practical applications.