A0892
Title: Estimation of process capability indices for high-dimensional data
Authors: Shuai Wang - Hainan University (China) [presenting]
Mengying You - Shanghai University of International Business and Economics (China)
Yi Li - Chengdu Technological University (China)
Jyun-You Chiang - Southwestern University of Finance and Economics (China)
Abstract: The purpose is to address the challenges in estimating multivariate process capability indices (MPCIs) for high-dimensional data in industrial processes. In modern manufacturing, processes are often characterized by multiple correlated quality characteristics, leading to complex dependencies in the data. However, traditional multivariate process capability index estimations are insufficient for these high-dimensional settings, as the estimation errors increase significantly. Furthermore, existing high-dimensional estimation methods do not guarantee that the estimated index falls within the correct value range. Therefore, the focus is on a framework for estimating high-dimensional MPCIs that modifies the adaptive thresholding estimation method by imposing a positive definite constraint, making it more suitable for multivariate process capability indices. Monte Carlo simulations are used to assess the performance of the proposed method under various sparsity levels and dimensionality, showing that it outperforms existing techniques, especially as dimensionality increases. Additionally, a real-world case study is presented from semiconductor manufacturing, demonstrating the practical applicability of the proposed method for high-dimensional process capability analysis. The results highlight the importance of accurate covariance matrix estimation and the need for tailored approaches to high-dimensional MPCIs analysis in modern industrial applications.
