Inference for the Process Performance Index of Products on the Basis of Power-Normal Distribution

The process performance index (PPI) can be a simple metric to connect the conforming rate of products. properties PPI have been well studied for normal distribution and other widely used lifetime distributions, such as Weibull, Gamma, Pareto distributions. Assume that quality characteristic product follows power-normal distribution. Statistical inference procedures are established. maximum likelihood estimation method model parameters is investigated exact Fisher information matrix derived. We discuss drawbacks using obtain confidence interval parameters. parametric bootstrap percentile bias-corrected methods proposed approximate intervals PPI. Monte Carlo simulations conducted evaluate methods. One example about flow width resist in hard-bake illustration.