Prediction Method of Mortar Compressive Strength Based on Feature Analysis and Deep Learning

In order to improve the prediction accuracy of concrete compressive strength, a prediction model of mortar concrete compressive strength based on convolutional neural network is proposed. Among them, the interpretable features related to the mortar concrete compressive strength are analyzed, 9 features are selected as input features of CNN convolutional neural network, and then the compressive strength is predicted by convolutional neural network. The experimental results show that the prediction model of convolutional neural network can achieve accurate prediction of the mortar concrete compressive strength and can accurately describe the change of compressive strength in various scenarios. Compared with the standard method, the prediction accuracy of concrete compressive strength by convolutional neural network is higher, reaching 9.63, 13.84 and 8.53 respectively on the three evaluation indexes of RMSE, MAPE and MAE. The above results show that the designed concrete compressive strength prediction model based on convolutional neural network can predict the concrete compressive strength with high precision and high reliability, and it can be applied to practical engineering scenarios.