基于随机森林的高性能混凝土抗压强度预测

Prediction of Compressive Strength of Concrete Based on Random Forests

抗压强度是混凝土重要参数之一,开展混凝土抗压强度预测对混凝土设计具有较高的参考价值.以多特征输入的高性能混凝土抗压强度为研究对象,构建1030个样本为数据集,结合随机森林算法对高性能混凝土抗压强度进行预测.此外,在同一数据集上还进行了支持向量回归分析和多层感知机预测.结果表明:三种机器学习模型中随机森林模型的预测精度最高,随机森林算法在测试集上的R^(2)得分0.902,MAE得分3.761,MAPE得分12.807,RMSE得分5.342;随着混凝土抗压强度增大,随机森林模型预测准确率逐渐提升,当混凝土抗压强度大于50 MPa时,随机森林的误差逐渐趋于收敛,预测精度也达到最高.

Compressive strength is one of the important parameters of concrete,so it has high reference value for concrete design to carry out concrete strength prediction.In this paper,the compressive strength of high performance concrete with multi feature input is taken as the research object,1030 samples are constructed as the data set,and the random forest algorithm is used to predict the compressive strength of high performance concrete.Besides,support vector regression analysis and multi layer perceptron prediction are carried out on the same data set.The results show that the random forest model's prediction accuracy is the highest among the three machine learning models.The R^(2) score of the random forest algorithm is 0.902,the MAE score is 3.761,the MAPE score is 12.807,and the RMSE score is 5.342;with the increase of concrete compressive strength,the prediction accuracy of Random forest model is gradually improved.When the concrete compressive strength is greater than 50 MPa,the error of random forest model tends to converge and the prediction accuracy is the highest.