摘要
为解决传统感性意象设计创新中真实的用户信息极少及需案例推理的问题,提出了基于网络评价数据的产品感性意象无偏差设计方法.首先,基于网络爬虫获取互联网中用户对产品外观的文本评价数据;接着,运用文本挖掘技术分析产品感性需求及意象信息,并构建词向量实现意象的参数化表达;然后,通过参数化曲线描绘目标产品的造型特征,建立产品形态特征的参数化数据;最后,基于最大信息系数选择感性意象参数的维度,并通过随机森林构建意象与设计特征间映射关系,实现对新产品设计特征参数及其取值范围的预测,明确产品感性意象创新设计参考依据.以三厢轿车侧轮廓设计为例,所建立的意象与设计特征间映射关系中,正确率在0.5~0.7之间的比例为86.67%,设计结果满足了主要的感性意象需求,表明通过该方法能够将用户真实的评价数据转化为意象设计求解过程中重要的输入信息,从而开展基于用户感性诉求描述的无案例推理的意象创新设计活动,高效实现感性意象无偏差设计.
To solve the problem of rare real user information and case-based reasoning in traditional kansei image design, an unbiased design method for product kansei image based on network evaluation data was proposed. Firstly, the network crawler was used to obtain the comment text of the product appearance in the Internet. Then, the text mining was introduced to analyze the product images and user demands for kansei, and the word vector was constructed to convert the kansei images into parameters. Furthermore, the morphological characteristics of the target product were depicted by the parametric curve, consequently, the parameterized data of product morphological characteristics were determined. Finally, the maximum information coefficient was calculated to select the dimensions of image parameters, and the mapping relationship based on random forest was constructed, the design parameters and its ranges of product were predicted based on mapping relationship. Taking the side profile design of the sedan car as an example, in the mapping relationship between the kansei image and design features, a correct ratio between 0.5 and 0.7 was 86.67%. The results show that the method can transform the user’s real evaluation data into important input information in the image design process, so as to carry out the image-innovative design activities based on the description of the user’s kansei demand without case-based reasoning. Thus, the unbiased design for images is efficiently realized.
作者
林丽
张云鹍
牛亚峰
阳明庆
Lin Li;Zhang Yunkun;Niu Yafeng;Yang Mingqing(Key Laboratory of Advanced Manufacturing Technology of Ministry of Education,Guizhou University,Guiyang 550025,China;School of Mechanical Engineering,Guizhou University,Guiyang 550025,China;School of Mechanical Engineering,Southeast University,Nanjing 211189,China)
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2020年第1期26-32,共7页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金资助项目(51465007,51865003)
贵州省科技计划资助项目(黔科合平台人才[2018]5781)
关键词
产品设计
感性工学
创新设计
用户知识
product design
kansei engineering
innovative design
user knowledge
