Analysis methodology for semiconductor yield by data mining

Tsuda Hidetaka, Hidehiro Shirai, Masahiro Terabe, Kazuo Hashimoto, Ayumi Shinohara

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

The conventional semiconductor yield analysis is a hypothesis verification process, which heavily depends on engineers' knowledge. Data mining methodology, on the other hand, is a hypothesis discovery process that is free from this constraint. This paper proposes a data mining method for semiconductor yield analysis, which consists of the following two phases: discovering hypothetical failure causes by regression tree analysis and verifying the hypotheses by visualizing the measured data based on engineers' knowledge. It is shown, through experiment under the real environment, that the proposed method detects hypothetical failure causes, which were considered practically impossible to detect, and that yield improvement is achieved by taking preventive actions based on the detected failure causes.

Original languageEnglish
JournalIEEJ Transactions on Industry Applications
Volume129
Issue number12
DOIs
Publication statusPublished - 2009
Externally publishedYes

Fingerprint

Data mining
Semiconductor materials
Engineers
Experiments

Keywords

  • Data mining
  • Hypothesis discovery
  • Regression tree analysis
  • Semiconductor
  • Yield analysis

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Industrial and Manufacturing Engineering

Cite this

Analysis methodology for semiconductor yield by data mining. / Hidetaka, Tsuda; Shirai, Hidehiro; Terabe, Masahiro; Hashimoto, Kazuo; Shinohara, Ayumi.

In: IEEJ Transactions on Industry Applications, Vol. 129, No. 12, 2009.

Research output: Contribution to journalArticle

Hidetaka, Tsuda ; Shirai, Hidehiro ; Terabe, Masahiro ; Hashimoto, Kazuo ; Shinohara, Ayumi. / Analysis methodology for semiconductor yield by data mining. In: IEEJ Transactions on Industry Applications. 2009 ; Vol. 129, No. 12.
@article{02434e8346694c54bee391394726b95a,
title = "Analysis methodology for semiconductor yield by data mining",
abstract = "The conventional semiconductor yield analysis is a hypothesis verification process, which heavily depends on engineers' knowledge. Data mining methodology, on the other hand, is a hypothesis discovery process that is free from this constraint. This paper proposes a data mining method for semiconductor yield analysis, which consists of the following two phases: discovering hypothetical failure causes by regression tree analysis and verifying the hypotheses by visualizing the measured data based on engineers' knowledge. It is shown, through experiment under the real environment, that the proposed method detects hypothetical failure causes, which were considered practically impossible to detect, and that yield improvement is achieved by taking preventive actions based on the detected failure causes.",
keywords = "Data mining, Hypothesis discovery, Regression tree analysis, Semiconductor, Yield analysis",
author = "Tsuda Hidetaka and Hidehiro Shirai and Masahiro Terabe and Kazuo Hashimoto and Ayumi Shinohara",
year = "2009",
doi = "10.1541/ieejias.129.1201",
language = "English",
volume = "129",
journal = "IEEJ Transactions on Industry Applications",
issn = "0913-6339",
publisher = "The Institute of Electrical Engineers of Japan",
number = "12",

}

TY - JOUR

T1 - Analysis methodology for semiconductor yield by data mining

AU - Hidetaka, Tsuda

AU - Shirai, Hidehiro

AU - Terabe, Masahiro

AU - Hashimoto, Kazuo

AU - Shinohara, Ayumi

PY - 2009

Y1 - 2009

N2 - The conventional semiconductor yield analysis is a hypothesis verification process, which heavily depends on engineers' knowledge. Data mining methodology, on the other hand, is a hypothesis discovery process that is free from this constraint. This paper proposes a data mining method for semiconductor yield analysis, which consists of the following two phases: discovering hypothetical failure causes by regression tree analysis and verifying the hypotheses by visualizing the measured data based on engineers' knowledge. It is shown, through experiment under the real environment, that the proposed method detects hypothetical failure causes, which were considered practically impossible to detect, and that yield improvement is achieved by taking preventive actions based on the detected failure causes.

AB - The conventional semiconductor yield analysis is a hypothesis verification process, which heavily depends on engineers' knowledge. Data mining methodology, on the other hand, is a hypothesis discovery process that is free from this constraint. This paper proposes a data mining method for semiconductor yield analysis, which consists of the following two phases: discovering hypothetical failure causes by regression tree analysis and verifying the hypotheses by visualizing the measured data based on engineers' knowledge. It is shown, through experiment under the real environment, that the proposed method detects hypothetical failure causes, which were considered practically impossible to detect, and that yield improvement is achieved by taking preventive actions based on the detected failure causes.

KW - Data mining

KW - Hypothesis discovery

KW - Regression tree analysis

KW - Semiconductor

KW - Yield analysis

UR - http://www.scopus.com/inward/record.url?scp=75349111143&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=75349111143&partnerID=8YFLogxK

U2 - 10.1541/ieejias.129.1201

DO - 10.1541/ieejias.129.1201

M3 - Article

AN - SCOPUS:75349111143

VL - 129

JO - IEEJ Transactions on Industry Applications

JF - IEEJ Transactions on Industry Applications

SN - 0913-6339

IS - 12

ER -