Abstract:

Information driven by uncertain user requirements usually has semantic, missing and redundant information. It can reduce accuracy of the mapping model from user requirements to product design parameters. In response to these problems, a fusion mapping model of fuzzy rough set+SVM was established. Fuzzy rough set was used to do granulation for quantified user data to fill missing information. The attribute reduction of rough set was used to eliminate redundant information and reduce space dimensions of input information. Then, SVM was used to construct the mapping model from user requirements to design parameters. Finally, the fusion model was applied to injection molding machine to verify its applicability.

Key words: uncertain, fuzzy rough set, SVM, mapping model

摘要：

不确定用户需求信息中存在语义、信息缺失和冗余现象,针对从用户需求信息到产品定制设计参数映射模型准确率低的问题,建立了模糊粗糙集+SVM的融合映射模型。利用模糊粗糙集对用户需求数据做粒化处理以补全缺失信息,再利用其属性约简消除冗余信息,降低空间输入维数。利用SVM构建用户需求信息到设计参数的映射模型。最后,通过注塑机产品实例分析验证了该融合模型的适用性。
不确定;模糊粗糙集;支持向量机;映射模型

关键词: 不确定, 模糊粗糙集, 支持向量机, 映射模型