Comparative Study on Data Mining Methods in Structural Reliability Prediction
碩士 === 國立臺灣科技大學 === 營建工程系 === 103 === The goal of reliability-based design optimization (RBDO) is to find the optimal structure design with minimum cost subjected to reliability constraint such as maximum failure probability limit. RBDO has two processes which are design optimization and reliability...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2015
|
Online Access: | http://ndltd.ncl.edu.tw/handle/73211235830561070436 |
id |
ndltd-TW-103NTUS5512070 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-103NTUS55120702017-01-07T04:08:46Z http://ndltd.ncl.edu.tw/handle/73211235830561070436 Comparative Study on Data Mining Methods in Structural Reliability Prediction Comparative Study on Data Mining Methods in Structural Reliability Prediction Willy Husada 何少悅 碩士 國立臺灣科技大學 營建工程系 103 The goal of reliability-based design optimization (RBDO) is to find the optimal structure design with minimum cost subjected to reliability constraint such as maximum failure probability limit. RBDO has two processes which are design optimization and reliability analysis. Since failure probability is usually small, it takes a large amount of computation time for accurate estimation in reliability analysis. Surrogate models are usually created to replace the time-consuming reliability analysis. In this empirical study, we use several data mining methods with focus on three methods, classification and regression tree (CART), artificial neural network (ANN) and support vector machine (SVM) to create the surrogate models on a empirical benchmark case. Data mining is used because it can find the hidden rules from a training data set and create a surrogate model based on its pattern recognition. In this study, we aim to find the best data mining method in predicting the failure probability in terms of prediction accuracy and computational efficiency which divided into two parts: classification and regression. The main findings of this study is that for one best setting, the ANN method performed better than CART and SVM in both classification and regression in term of prediction accuracy. But, the CART method is more stable in terms of accuracy range. Moreover, the computation time of the CART method is much shorter and therefore superior to both ANN and SVM. In general, the CART method is more favourable than the ANN and SVM methods since it is very efficient in terms of computation time and attain high prediction accuracy. I-Tung Yang 楊亦東 2015 學位論文 ; thesis 167 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立臺灣科技大學 === 營建工程系 === 103 === The goal of reliability-based design optimization (RBDO) is to find the optimal structure design with minimum cost subjected to reliability constraint such as maximum failure probability limit. RBDO has two processes which are design optimization and reliability analysis. Since failure probability is usually small, it takes a large amount of computation time for accurate estimation in reliability analysis. Surrogate models are usually created to replace the time-consuming reliability analysis. In this empirical study, we use several data mining methods with focus on three methods, classification and regression tree (CART), artificial neural network (ANN) and support vector machine (SVM) to create the surrogate models on a empirical benchmark case. Data mining is used because it can find the hidden rules from a training data set and create a surrogate model based on its pattern recognition. In this study, we aim to find the best data mining method in predicting the failure probability in terms of prediction accuracy and computational efficiency which divided into two parts: classification and regression. The main findings of this study is that for one best setting, the ANN method performed better than CART and SVM in both classification and regression in term of prediction accuracy. But, the CART method is more stable in terms of accuracy range. Moreover, the computation time of the CART method is much shorter and therefore superior to both ANN and SVM. In general, the CART method is more favourable than the ANN and SVM methods since it is very efficient in terms of computation time and attain high prediction accuracy.
|
author2 |
I-Tung Yang |
author_facet |
I-Tung Yang Willy Husada 何少悅 |
author |
Willy Husada 何少悅 |
spellingShingle |
Willy Husada 何少悅 Comparative Study on Data Mining Methods in Structural Reliability Prediction |
author_sort |
Willy Husada |
title |
Comparative Study on Data Mining Methods in Structural Reliability Prediction |
title_short |
Comparative Study on Data Mining Methods in Structural Reliability Prediction |
title_full |
Comparative Study on Data Mining Methods in Structural Reliability Prediction |
title_fullStr |
Comparative Study on Data Mining Methods in Structural Reliability Prediction |
title_full_unstemmed |
Comparative Study on Data Mining Methods in Structural Reliability Prediction |
title_sort |
comparative study on data mining methods in structural reliability prediction |
publishDate |
2015 |
url |
http://ndltd.ncl.edu.tw/handle/73211235830561070436 |
work_keys_str_mv |
AT willyhusada comparativestudyondataminingmethodsinstructuralreliabilityprediction AT héshǎoyuè comparativestudyondataminingmethodsinstructuralreliabilityprediction |
_version_ |
1718407185827889152 |