Using an attribute conversion approach for sample generation to learn small data with highly uncertain features

Der-Chiang Li, Qi Shi Shi, Ming Da Li

Research output: Contribution to journalArticle

Abstract

Accelerating new product development has become an important marketing strategy for manufacturers who are competing globally. However, this may lead to the small data learning issue. Although machine learning algorithms are used to extract knowledge from training samples, algorithms may not output satisfactory predictions when training sizes are small. This paper provides a real case of a TFT-LCD (thin film transistor liquid crystal display) maker when a new strengthened cover glass is developed using chemical processes. With very little prior experience about the processes involved, engineers attempted to improve the yield rates by determining the parameters from a few pilot-run data. However, owing to the fact that the processes were different from those required to make TFT-LCD panels, the highly uncertain characteristics of the processes led to the use of two virtual sample generation (VSG) approaches, bootstrap aggregating (bagging) and the synthetic minority over-sampling technique, from which unsatisfactory results were obtained. Accordingly, this study was used to develop a systematic VSG method based on fuzzy theory to tackle the learning issue. The experimental results show that support vector regressions built with training sets containing the proposed samples present more precise predictions and thus can help engineers infer more correct manufacturing parameters.

Original languageEnglish
Pages (from-to)4954-4967
Number of pages14
JournalInternational Journal of Production Research
Volume56
Issue number14
DOIs
Publication statusPublished - 2018 Jul 18

Fingerprint

Thin film transistors
Liquid crystal displays
Engineers
Product development
Learning algorithms
Learning systems
Marketing
Sampling
Glass
Prediction

All Science Journal Classification (ASJC) codes

  • Strategy and Management
  • Management Science and Operations Research
  • Industrial and Manufacturing Engineering

Cite this

@article{7a4f21ebeb5845e79bd3f3cda8940133,
title = "Using an attribute conversion approach for sample generation to learn small data with highly uncertain features",
abstract = "Accelerating new product development has become an important marketing strategy for manufacturers who are competing globally. However, this may lead to the small data learning issue. Although machine learning algorithms are used to extract knowledge from training samples, algorithms may not output satisfactory predictions when training sizes are small. This paper provides a real case of a TFT-LCD (thin film transistor liquid crystal display) maker when a new strengthened cover glass is developed using chemical processes. With very little prior experience about the processes involved, engineers attempted to improve the yield rates by determining the parameters from a few pilot-run data. However, owing to the fact that the processes were different from those required to make TFT-LCD panels, the highly uncertain characteristics of the processes led to the use of two virtual sample generation (VSG) approaches, bootstrap aggregating (bagging) and the synthetic minority over-sampling technique, from which unsatisfactory results were obtained. Accordingly, this study was used to develop a systematic VSG method based on fuzzy theory to tackle the learning issue. The experimental results show that support vector regressions built with training sets containing the proposed samples present more precise predictions and thus can help engineers infer more correct manufacturing parameters.",
author = "Der-Chiang Li and Shi, {Qi Shi} and Li, {Ming Da}",
year = "2018",
month = "7",
day = "18",
doi = "10.1080/00207543.2018.1444813",
language = "English",
volume = "56",
pages = "4954--4967",
journal = "International Journal of Production Research",
issn = "0020-7543",
publisher = "Taylor and Francis Ltd.",
number = "14",

}

Using an attribute conversion approach for sample generation to learn small data with highly uncertain features. / Li, Der-Chiang; Shi, Qi Shi; Li, Ming Da.

In: International Journal of Production Research, Vol. 56, No. 14, 18.07.2018, p. 4954-4967.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Using an attribute conversion approach for sample generation to learn small data with highly uncertain features

AU - Li, Der-Chiang

AU - Shi, Qi Shi

AU - Li, Ming Da

PY - 2018/7/18

Y1 - 2018/7/18

N2 - Accelerating new product development has become an important marketing strategy for manufacturers who are competing globally. However, this may lead to the small data learning issue. Although machine learning algorithms are used to extract knowledge from training samples, algorithms may not output satisfactory predictions when training sizes are small. This paper provides a real case of a TFT-LCD (thin film transistor liquid crystal display) maker when a new strengthened cover glass is developed using chemical processes. With very little prior experience about the processes involved, engineers attempted to improve the yield rates by determining the parameters from a few pilot-run data. However, owing to the fact that the processes were different from those required to make TFT-LCD panels, the highly uncertain characteristics of the processes led to the use of two virtual sample generation (VSG) approaches, bootstrap aggregating (bagging) and the synthetic minority over-sampling technique, from which unsatisfactory results were obtained. Accordingly, this study was used to develop a systematic VSG method based on fuzzy theory to tackle the learning issue. The experimental results show that support vector regressions built with training sets containing the proposed samples present more precise predictions and thus can help engineers infer more correct manufacturing parameters.

AB - Accelerating new product development has become an important marketing strategy for manufacturers who are competing globally. However, this may lead to the small data learning issue. Although machine learning algorithms are used to extract knowledge from training samples, algorithms may not output satisfactory predictions when training sizes are small. This paper provides a real case of a TFT-LCD (thin film transistor liquid crystal display) maker when a new strengthened cover glass is developed using chemical processes. With very little prior experience about the processes involved, engineers attempted to improve the yield rates by determining the parameters from a few pilot-run data. However, owing to the fact that the processes were different from those required to make TFT-LCD panels, the highly uncertain characteristics of the processes led to the use of two virtual sample generation (VSG) approaches, bootstrap aggregating (bagging) and the synthetic minority over-sampling technique, from which unsatisfactory results were obtained. Accordingly, this study was used to develop a systematic VSG method based on fuzzy theory to tackle the learning issue. The experimental results show that support vector regressions built with training sets containing the proposed samples present more precise predictions and thus can help engineers infer more correct manufacturing parameters.

UR - http://www.scopus.com/inward/record.url?scp=85042919496&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042919496&partnerID=8YFLogxK

U2 - 10.1080/00207543.2018.1444813

DO - 10.1080/00207543.2018.1444813

M3 - Article

VL - 56

SP - 4954

EP - 4967

JO - International Journal of Production Research

JF - International Journal of Production Research

SN - 0020-7543

IS - 14

ER -