Dynamic Modulus Prediction of a High-Modulus Asphalt Mixture

Dynamic modulus is a key evaluation index of the high-modulus asphalt mixture, but it relatively difficult to test and collect its data. The purpose achieve accurate prediction dynamic mixture further optimize design process mixture. Five high-temperature performance indexes were selected. correlation between above five was analyzed. On this basis, models based on small sample data established by multiple regression, general regression neural network (GRNN), support vector machine (SVM) network. According parameter adjustment cross-validation, output stability accuracy different compared evaluated. most effective model recommended. results show that SVM has more significant than GRNN model. Its error 0.98–9.71%. Compared with other two models, declined 0.50–11.96% 3.76–13.44%. recommended as