Modeling manufacturing processes using fuzzy regression with the detection of outliers

Chun Kit Kwong, Y. Chen, H. Wong

Research output: Journal article publicationJournal articleAcademic researchpeer-review

10 Citations (Scopus)

Abstract

Empirical modeling, which involves various common techniques such as statistical regression, artificial neural networks and fuzzy logic modeling, is a popular approach to developing models for manufacturing processes. Among those techniques, statistical regression is the most popular one used to develop the explicit type of empirical models. However, if the experimental data and results contain a substantial degree of fuzziness, fuzzy regression is more appropriate for use in developing empirical models based on such data and results. In recent years, attempts have been made to use fuzzy regression to model manufacturing processes. However, it has been recognized that the existence of outliers can have a great effect on the prediction accuracy of a fuzzy regression model. This problem has not been well addressed in the previous studies on fuzzy regression. In this paper, an algorithm for detecting outliers based on Peters' fuzzy regression is proposed. The application of the algorithm to developing a fuzzy regression-based process model of the dispensing of fluid for IC chip encapsulation is described. Finally, the results of the validation of the models are discussed.
Original languageEnglish
Pages (from-to)547-557
Number of pages11
JournalInternational Journal of Advanced Manufacturing Technology
Volume36
Issue number5-6
DOIs
Publication statusPublished - 1 Mar 2008

Keywords

  • Fluid dispensing
  • Fuzzy regression
  • Outlier detection
  • Process modeling

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Software
  • Mechanical Engineering
  • Computer Science Applications
  • Industrial and Manufacturing Engineering

Cite this