Concrete Strength Prediction Using Linear Regression of Machine Learning Algorithm


  • Peggy Suenie Anak Achong Student
  • Nickholas Anting Anak Guntor Lecturer of UTHM


Compressive Strength, Prediction, Regression, Performance, R-squared


Compressive strength is a performance measurement to determine the quality of the concrete. It is one of the crucial parameters used by industries to evaluate the performance of concrete. This study will be more focused on the machine learning model. The Multiple Linear Regression (MLR) algorithm is used to train the machine learning. The model was developed using KNIME analytical platform software. The data set consists of 202 observations with 26 attributes. The dataset was divided into 80% for training set and 20% for testing set. The training set is used to train the algorithm model, while the testing set is used to evaluate the performance of the model. A comparison was made between the MLR model and the Polynomial Regression (PR) model. Fine aggregate (kg/m3) was found to be the most influential contributor to increasing the compressive strength of concrete in the MLR model. The MLR model's R-squared value is 0.589, lower than the R-squared PR model, which is 0.745. The performance of the models was eventually indicated by using the Linear Correlation node. It is clearly stated indicated that both models are considered good models for the prediction in the machine learning model. Initially, the MLR model was proposed to be used as the machine learning prediction model. Compared with the PR model, the R-squared value of the MLR model was lower than PR model; but, it still indicates the model can still produce a good model. Thus, the PR model is recommended in the machine learning prediction model to predict the concrete compressive strength.


Keywords: Compressive strength, Prediction, Regression, Performance, R-squared




How to Cite

Achong, P. S. A., & Guntor, N. A. A. (2021). Concrete Strength Prediction Using Linear Regression of Machine Learning Algorithm. Recent Trends in Civil Engineering and Built Environment, 2(1), 691–699. Retrieved from