中文摘要：

为满足人们对高速列车造型的感性意象需求,提出基于文本挖掘与神经网络的高速列车意象造型设计方法。采用文本挖掘和模块分解,确定高速列车的感性意象集和设计模块库;采用群决策聚类的方法集结感性意象集的评分,从而提高调查样本的信度;应用BP神经网络,获得高速列车设计模块库和感性意象集之间的映射关系。基于建立的映射关系,选择设计模块进行组合与调整,形成符合不同感性意象需求的设计方案。结果表明：该方法较好地模拟了设计思维,可为高速列车造型设计提供有效的辅助与支持。

英文摘要：

In order to meet the requirements of user perception image for high-speed train modeling, an image modeling design method of high-speed train was presented. The perception image set and design module library were built using text mining and modular decomposition. The scores of perception image were calculated using the cluster method of group decision to improve the reliability of the samples. The mapping relationship between the perception image set and design module library of high-speed train was obtained using BP neural network. The design modules were selected to perform combination and adjustment based on the built mapping relationship. The design schemes for different perception images were put forward. The results showed that this method could simulate the design thinking effectively and provide guide and support for the high-speed train modeling design effectively.