论文部分内容阅读
为了直接反映可控边界参数与热耗率的映射关系,基于υ-SVM建立了可控边界参数与热耗率的回归模型,选取与热耗率关联性强的可控边界参数作为输入参数,并应用灰色关联度模型进行验证,详细地描述了基于Libsvm软件建立υ-SVM回归模型的过程,并与BP神经网络模型进行对比.结果表明:在小样本情况下,υ-SVM模型回归精度更高,具有更好的泛化能力;在输入参数小幅波动的情况下,υ-SVM模型的输出结果基本稳定,具有很好的鲁棒性,满足实际应用的精度要求.
In order to directly reflect the mapping relationship between controllable boundary parameters and heat consumption rate, a regression model of controllable boundary parameters and heat consumption rate was established based on υ-SVM. The controllable boundary parameters with strong correlation with heat rate were chosen as input parameters, And the gray relational degree model is used for verification.The process of establishing υ-SVM regression model based on Libsvm software is described in detail and compared with BP neural network model.The results show that the regression accuracy of υ-SVM model is more High, and has better generalization ability. Under the condition of slight fluctuation of input parameters, the output of υ-SVM model is basically stable and has good robustness to meet the precision requirements of practical application.