Back to Search
Start Over
Key Process Variable Identification for Quality Classification Based on PLSR Model and Wrapper Feature Selection
- Source :
- Proceedings of 2012 3rd International Asia Conference on Industrial Engineering and Management Innovation (IEMI2012) ISBN: 9783642330117
- Publication Year :
- 2012
- Publisher :
- Springer Berlin Heidelberg, 2012.
-
Abstract
- In modern manufacturing, hundreds of process variables are collected, and it is usually difficult to identify the most informative ones. Partial Least Square Regression provides an efficient way to evaluate each variable, but it cannot evaluate any variable subset as a whole. In the paper, a new framework of key process variable identification is proposed. It combines PLSR model and wrapper feature selection to firstly assess every variable individually and then the top variables in groups. Five datasets are tested, and the average classification accuracy is higher and the key process variables identified are less than the available approaches.
- Subjects :
- business.industry
Computer science
media_common.quotation_subject
Process (computing)
Pattern recognition
Feature selection
Process variable
computer.software_genre
Identification (information)
Variable (computer science)
Partial least squares regression
Key (cryptography)
Quality (business)
Artificial intelligence
Data mining
business
computer
media_common
Subjects
Details
- ISBN :
- 978-3-642-33011-7
- ISBNs :
- 9783642330117
- Database :
- OpenAIRE
- Journal :
- Proceedings of 2012 3rd International Asia Conference on Industrial Engineering and Management Innovation (IEMI2012) ISBN: 9783642330117
- Accession number :
- edsair.doi...........a2a70d1b479956ca1d3ffa0ab2e69613