MLPCA based logistical regression analysis for pattern clustering in manufacturing processes

Feng Zhang*, Daniel Apley

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


This article proposes a method of improving pattern clustering accuracy combined with a logistical regression model for manufacturing processes with binary inspection outputs. A latent variable model was incorporated into the classical logistical regression model, involving the use of Maximum Likelihood Principal Component Analysis (MLPCA) to identify the underlying variation sources governing the behaviors of the high dimensional measurable variables. The highly correlated continuous measurable predictors are projected onto a lower dimensional latent space, followed by a more precise pattern clustering algorithm to the inspected manufacturing products for in-line process monitoring. The example of the visual inspects from semiconductor manufacturing processes is shown that this new pattern clustering algorithm could help identify the root causes of the variations.

Original languageEnglish (US)
Pages (from-to)898-902
Number of pages5
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
StatePublished - 2004

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'MLPCA based logistical regression analysis for pattern clustering in manufacturing processes'. Together they form a unique fingerprint.

Cite this