基于支持向量机的产品感性意象值预测方法

为设计出符合消费者感性需求的产品,提出基于支持向量机的产品感性意象值预测方法。先确定产品的感性意象、造型设计要素以及感性评价矩阵。在此基础上,以造型设计要素为自变量,以感性意象评价值为因变量,利用LIBSVM软件,通过对惩罚函数、不敏感损失函数以及核函数等相关参数的分析设置,建立产品感性意象值的预测模型。结合办公座椅进行研究,结果表明支持向量机具有较高的预测精度,所提出的方法是正确可行的。

Prediction Method of Product Kansei Image Value Based on Support Vector Machine

In order to design products that meet consumers′ affective needs,prediction method of product kansei image value based on support vector machine was proposed.Firstly,product kansei images,form design elements,and kansei evaluation matrix were determined.Based on this,taking the form design elements as independent variables and the kansei evaluation values as dependent variables,the prediction model of product kansei image value was built using LIBSVM software through analyzing and setting the parameters of penalty function,insensitive loss function and nuclear function.An experimental study of office chair was conducted,and the results suggested that support vector machine had a better prediction performance,and the presented method was valid and feasible.