Advanced Multivariate Outlier Detection for Automotive Road-to-Zero Defect Strategy

Reliability becomes incredibly critical and the quality goal has changed from 10 DPPM (defective parts per million) to 10 DPPB (defective parts per billion) because of the significant increase of semiconductor electronic components in automotive vehicles. This paper creates Python programs with principal component analysis and multivariate outlier detection models that identify outliers correlated to latent failures post reliability stress failures or customer returns.