复杂产品的关键质量特性识别

Identification of Critical-to-Quality for Complex Products

针对复杂产品关键质量特征特性识别中存在的数据高维度、小样本、不平衡问题,提出一种基于Bootstrap K-split Lasso(B-K-Lasso)的关键质量特征识别方法,并采用UCI数据库中的SECOM数据集进行仿真实验。结果表明:该方法能够有效识别复杂产品的关键质量特性,并且对不合格品的识别率显著提升。

In order to solve the problems of high dimensions, small sample size and unbalanced data in the recognition of critical-to-quanlity of complex products, this paper proposes a critical-to-quanlity recognition method based on Bootstrap K-split Lasso(B-K-split Lasso). The simulation results of SECOM data set in UCI database show that this method can effectively identify the critical-to-quanlity of complex products and significantly improve the accuracy of nonconforming products.