Abstract

Concrete compressive strength is a critical property that determines the structural performance and durability of construction materials. Traditionally, compressive strength is measured through laboratory testing procedures that require curing periods of several days or weeks. These tests are time consuming and expensive, especially when large numbers of samples are required. Machine learning techniques provide an efficient alternative for predicting compressive strength based on concrete mixture proportions. This study investigates the effectiveness of five machine learning models for predicting the compressive strength of concrete using a synthetic dataset representing common mix design parameters. The models used in this study include Linear Regression, Decision Tree Regression, Random Forest Regression, Support Vector Regression, and Artificial Neural Networks. The dataset contains variables such as cement content, blast furnace slag, fly ash, and water, and superplasticizer, coarse aggregate, fine aggregate, and curing age. The models were trained and evaluated using a standard training and testing approach. Experimental results demonstrate that ensemble learning models such as Random Forest achieve higher predictive accuracy compared to traditional regression techniques. The study highlights the potential of machine learning techniques for improving efficiency in concrete strength prediction and assisting engineers in optimizing mix designs.