Predicting Confined Compressive Strength of Concrete Using Machine Learning Approach

The main objective of this investigation is to develop a machine learning model, namely random forest model (RF), for estimating the confined compressive strength of concrete. In this regard, a total of 144 experimental results on the confined compressive strength of concrete were gathered from the published document sand used to develop the RF model. The results of the RF model were compared with those of the normal regression model, namely multiple linear regression (MLR). Based on the comparison, it is emphasized that the developed RF model in this article can predict the confined compressive strength of concrete more accurately than the MLR model. Also, the effect of input factors on the confined compressive strength of concrete estimation was explored using the sensitivity assessment. The results shown that the two parameters, comprising compressive strength of concrete core and diameter and thickness of tube ratio were found to be substantially important to forecast the confined compressive strength of concrete.