A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering
碩士 === 國立雲林科技大學 === 工業工程與管理系 === 105 === Control charts are the most common tool to monitor process, they can be used to assess whether the process is in control or not. Control charts detect signals that is not in control when variations of the process occur. However, control chart would detect o...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2017
|
Online Access: | http://ndltd.ncl.edu.tw/handle/37b28k |
id |
ndltd-TW-105YUNT0031043 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-105YUNT00310432018-05-13T04:29:21Z http://ndltd.ncl.edu.tw/handle/37b28k A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering 結合支撐向量機與模糊統計集群法於變動抽樣管制圖估計改變點之研究 LIOU, FU-JAN 劉富展 碩士 國立雲林科技大學 工業工程與管理系 105 Control charts are the most common tool to monitor process, they can be used to assess whether the process is in control or not. Control charts detect signals that is not in control when variations of the process occur. However, control chart would detect out of control signal with large amount of delay in most cases. To address this problem, many supplementary techniques are applied to control charts aiming to identify the exact change time in process. This study proposes a hybrid method to estimate the time of change in variable sampling X ̅ control chart, we assume the change type and the magnitude of process variations are unknown. For identifying the change type in process, this study used the support vector machine which would recognize control chart patterns to address the unknown change type. After we identify the change type, the estimation of the change time is accomplished by the Fuzzy statistical clustering. The study then conducts extended simulations to evaluate this hybrid method’s performance of change point estimation with different variable sampling strategies and different change types in process. The result show that the performance of estimating change point is very close to the real change point for upward and downward step change. On the other hand, the performance of estimating change point is later than the real change point for increasing and decreasing trend, and the larger the magnitude of disturbance change is, the better the performance of estimating change point is. The performance of proposed method is close to the performance of the change type be known, both of the performances of estimators are excellent. TORNG, CHAU-CHEN 童超塵 2017 學位論文 ; thesis 85 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立雲林科技大學 === 工業工程與管理系 === 105 === Control charts are the most common tool to monitor process, they can be used to assess whether the process is in control or not. Control charts detect signals that is not in control when variations of the process occur. However, control chart would detect out of control signal with large amount of delay in most cases. To address this problem, many supplementary techniques are applied to control charts aiming to identify the exact change time in process. This study proposes a hybrid method to estimate the time of change in variable sampling X ̅ control chart, we assume the change type and the magnitude of process variations are unknown. For identifying the change type in process, this study used the support vector machine which would recognize control chart patterns to address the unknown change type. After we identify the change type, the estimation of the change time is accomplished by the Fuzzy statistical clustering. The study then conducts extended simulations to evaluate this hybrid method’s performance of change point estimation with different variable sampling strategies and different change types in process. The result show that the performance of estimating change point is very close to the real change point for upward and downward step change. On the other hand, the performance of estimating change point is later than the real change point for increasing and decreasing trend, and the larger the magnitude of disturbance change is, the better the performance of estimating change point is. The performance of proposed method is close to the performance of the change type be known, both of the performances of estimators are excellent.
|
author2 |
TORNG, CHAU-CHEN |
author_facet |
TORNG, CHAU-CHEN LIOU, FU-JAN 劉富展 |
author |
LIOU, FU-JAN 劉富展 |
spellingShingle |
LIOU, FU-JAN 劉富展 A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering |
author_sort |
LIOU, FU-JAN |
title |
A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering |
title_short |
A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering |
title_full |
A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering |
title_fullStr |
A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering |
title_full_unstemmed |
A Hybrid Method for Estimating the Time of Change in Variable Sampling Control Chart Using Support Vector Machine and Fuzzy Statistical Clustering |
title_sort |
hybrid method for estimating the time of change in variable sampling control chart using support vector machine and fuzzy statistical clustering |
publishDate |
2017 |
url |
http://ndltd.ncl.edu.tw/handle/37b28k |
work_keys_str_mv |
AT lioufujan ahybridmethodforestimatingthetimeofchangeinvariablesamplingcontrolchartusingsupportvectormachineandfuzzystatisticalclustering AT liúfùzhǎn ahybridmethodforestimatingthetimeofchangeinvariablesamplingcontrolchartusingsupportvectormachineandfuzzystatisticalclustering AT lioufujan jiéhézhīchēngxiàngliàngjīyǔmóhútǒngjìjíqúnfǎyúbiàndòngchōuyàngguǎnzhìtúgūjìgǎibiàndiǎnzhīyánjiū AT liúfùzhǎn jiéhézhīchēngxiàngliàngjīyǔmóhútǒngjìjíqúnfǎyúbiàndòngchōuyàngguǎnzhìtúgūjìgǎibiàndiǎnzhīyánjiū AT lioufujan hybridmethodforestimatingthetimeofchangeinvariablesamplingcontrolchartusingsupportvectormachineandfuzzystatisticalclustering AT liúfùzhǎn hybridmethodforestimatingthetimeofchangeinvariablesamplingcontrolchartusingsupportvectormachineandfuzzystatisticalclustering |
_version_ |
1718638747108507648 |