Design of an Ensemble learning Model for improving Concrete Strength prediction efficiency of RC Structures via Multiparametric Analysis. (EMSRSMA)

Abstract

Prediction of Concrete Strength for reinforced concrete (RC) structures is a multimodal task that involves processing of a wide variety of strength parameter sets. Existing prediction models either showcase low efficiency or cannot be scaled for heterogenous construction sites. To overcome these issues, this paper proposes an ensemble learning model to improve the efficiency of evaluating concrete compressive strength using multiparametric analysis. The model combines multiple algorithms, including Naïve Bayes (NB), Multilayer Perceptron (MLP), k Nearest Neighbours (kNN), Support Vector Machines (SVM), and Logistic Regression (LR), to provide accurate and reliable predictions of compressive strength levels. The proposed ensemble approach leverages a diverse set of features, including cement, water, and aggregate content, curing time, and age, to achieve high prediction performance for heterogeneous construction sites. The proposed model's performance was evaluated using a comprehensive dataset of concrete mixtures, and the results show that the ensemble approach outperforms individual algorithms and achieves a higher level of accuracy, precision & recall when compared with existing techniques. The proposed model's success demonstrates the potential of ensemble learning methods for Concrete Strength prediction and provides a promising solution for improving the efficiency of construction material evaluation under real-time scenarios. The proposed model can assist in reducing costs and enhancing the reliability of construction projects by providing a more accurate assessment of Concrete Strength levels.