Neural Network Prediction and Sensitivity Analysis of Concrete Compressive Strength: Sensitivity and Feature Importance Analysis

In engineering practice, the measurement of concrete compressive strength is often very troublesome, time-consuming and laborious, so a method that can predict the strength of concrete is very important, this paper designs a BP neural network based on the prediction of the compressive strength of concrete, by means of the neural network constantly training for the prediction of the compressive strength of concrete, the prediction results can be reasoned, and the results of the actual value of the comparison of error The prediction results are rationalized, and the error of comparison with the actual value results is small, and the accuracy is high. At the same time, this paper calibrates the parameters of the neural network when predicting the compressive strength of concrete, which provides guidance for the selection of super-parameters. At the same time, this paper analyzes the importance of the features when the concrete is condensed and hardened, and obtains the importance index of each feature.