DocumentCode
539445
Title
A trend mining method for yield improvement based on trend in time series
Author
Tsuda, Hidetaka ; Shirai, Hidehiro ; Terabe, Masahiro ; Hashimoto, Kazuo ; Shinohara, Ayumi
Author_Institution
Fujitsu LSI Technology Limited, KSP R&D 3-2-1 Sakado, Takatsuku Kawasaki 213-0012, JAPAN
fYear
2008
fDate
27-29 Oct. 2008
Firstpage
247
Lastpage
250
Abstract
We have developed a trend mining method for yield improvement which extracts non-random trend in time series buried in huge time series data. This method efficiently extracts failure causes based on trend in time series that had been almost unable to be analyzed by conventional methods.
Keywords
Data analysis; Data mining; Feature extraction; Fluctuations; Manufacturing; Regression tree analysis; Time series analysis;
fLanguage
English
Publisher
ieee
Conference_Titel
Semiconductor Manufacturing (ISSM), 2008 International Symposium on
Conference_Location
Tokyo, Japan
ISSN
1523-553X
Electronic_ISBN
1523-553X
Type
conf
Filename
5714967
Link To Document