High correlated variables creator machine: Prediction of the compressive strength of concrete

In this paper, we introduce a novel hybrid model for predicting the compressive strength of concrete using Ultrasonic Pulse Velocity (UPV) and Rebound Number (RN). First, we collect 516 datasets from 8 studies of UPV and Rebound Hammer (RH) tests. Then, we propose the High Correlated Variables Creator Machine (HCVCM) to create the new variables that have a better correlation with the output in order to improve the prediction models. Three single models, including a Step-By-Step Regression (SBSR), Gene Expression Programming (GEP), an Adaptive Neuro-Fuzzy Inference System (ANFIS) as well as three hybrid models, i.e. HCVCM-SBSR, HCVCM-GEP, and HCVCM-ANFIS are employed to predict the compressive strength of concrete. The statistical parameters and error terms such as the coefficient of determination, the Root Mean Square Error (RMSE), Normalized Mean Square Error (NMSE), fractional bias, the maximum positive and negative errors, and the Mean Absolute Percentage Error (MAPE) are computed to evaluate the models. The results show that HCVCM-ANFIS can predict the compressive strength of concrete better than all other models. HCVCM improves the accuracy of ANFIS by 5% in the coefficient of determination, 10% in the RMSE, 3% in the NMSE, 20% in MAPE, and 7% in the maximum negative error. (C) 2021 Elsevier Ltd. All rights reserved.

Automated summary

This paper introduces a novel hybrid model to predict the compressive strength of concrete using HCVCM method, and demonstrates that the HCVCM-ANFIS model can predict the compressive strength more accurately than other models.