Forecasting and Understanding Compressive Properties of Concrete Using the Ensemble Learning Methods
作者
Mohit,L. Balwinder
标识
DOI:10.1109/icseiet58677.2023.10303444
摘要
Concrete is the backbone of modern-day architecture. The composition of concrete changes varies geologically. The compressive strength of concrete is an important factor which should be consider in order to test for the reliability of the concrete sample. The machine learning ensemble models present a way to predict values for compressive strength. The data is collected by conducting experiments for the study and the values of the study are fed into the database for building and testing of the ensemble models. The testing of the learning models is done using statistical metrics like mean absolute error, mean squared error, root mean squared error, and coefficient of determination. The different algorithms like random forest, k- nearest neighbors, and support vector regression were studied and used in order to design ensemble models. In the study, stacking-based regression with support vector regression, random forest and k- nearest neighbor with k- nearest neighbor as final estimator outperformed other models including bagging and random forest with coefficient of determination of 0.98. The trained machine learning models were also used to develop a web application for ease of use.