Interval Estimation of Process Capability Indices Based on the Quality Data of Supplied Products

Abstract

Process capability indices (PCIs) are employed to select an appropriate supplier for customers. However, data fraud events may lead to incorrect decisions on quality management, it is more reliable to assess the manufacturing process capability based on supplied products. According to the quality requirement of the specification limits, the quality data of supplied products are doubly truncated. Thus, process capability analysis from supplied products is adopted to evaluate the manufacturing process capability, supervise and check the quality data provided by suppliers. For this purpose, under normal assumption, EM algorithm is firstly adopted to estimate the unknown parameters of the quality distribution of supplied products. Then, a quantile-filling algorithm is proposed to convert the truncated data into the pseudo-complete data. Next, based on the pseudo-complete data, the interval estimation of PCIs is carried out by the generalized confidence interval method. Finally, a practical example is given to show the implement of the proposed method.