基于最小二乘支持向量机的QFD技术特性权重预测  

Weight prediction for technical characteristics in QFD based on LS-SVM

作　　者：陆佳圆[1] 谭建荣[1] 冯毅雄[1] 安相华[1] 

机构地区：[1]浙江大学流体传动及控制国家重点实验室,浙江杭州310027

出　　处：《机械设计》2011年第7期1-7,共7页Journal of Machine Design

基　　金：国家自然科学基金资助项目(50835008;50875237);国家科技重大专项资助项目(2009ZX04014-031)

摘　　要：针对质量功能展开(Quality Funcition Deployment,QFD)中技术特性权重的动态性,提出了预测其权重值的有效方法。该方法首先在获取顾客需求最终权重的基础上,通过顾客需求与技术特性之间的质量屋关联矩阵把顾客需求权重转化为技术特性权重;然后周期性地获取技术特性权重以构建权重数据的时间序列,并引入最小二乘支持向量机回归模型对权重数据进行处理,得到权重预测值;最后构建技术特性动态变化趋势分析图,以便于为产品的改进设计提供决策。以超高速电梯产品为例,对该方法的可行性和有效性进行具体说明。Aiming at the dynamic property of technical characteristics（TC） weight in Quality Function Deployment（QFD）,an effective method for predicting its weight value was presented.First,on the basis of the final weight required by customer,the customer requirement weight was transformed into technical characteristics weight by using the quality house incidence matrix between customer requirement and technical characteristics.Then,the time series of TC weight data was constructed by obtaining the TC weight periodically;the weight data was processed to calculate the predicted weight by means of introducing the Least Squares Support Vector Machines regression model.Finally,the dynamic variation trend analysis chart of TC weight was established so as to provide the decision for design development of products.A case study of super high speed elevator was provided to prove the feasibility and effectiveness of the proposed method in details.

关 键 词：质量功能展开 最小二乘支持向量机 技术特性 权重预测 

分 类 号：N94[自然科学总论—系统科学] TP391[自动化与计算机技术—计算机应用技术]