基于BP神经网络算法的橡胶配方性能预测软件

利用Matlab神经网络工具箱、图形用户界面环境以及BP神经网络开发出橡胶配方性能预测软件。软件具有友好的图形用户界面，便于输入参数及查看分析结果，并编译成可脱离Matlab环境运行的应用程序。该软件具备新神经网络的训练和保存功能，可在使用中不断升级完善。通过实例对软件的预测结果进行验证，证实了预测结果的准确性。该软件有很强的可移植性，使用方便，具有很好的工程应用价值。;Using Matlab neural network toolbox，graphical user interface and BP neural network，the performance prediction software of rubber formulation was developed.A friendly graphical user interface was designed to input the relevant parameters and output analysis results easily.This software can run independently without Matlab environment.It had the training and preserving functions of new neural network，and could be upgraded continuously in use.The validity of its predicted results was verified by a group of test data.This software possessed strong portability，and was convenient to use and had good engineering application value