Reliability assessment using probabilistic support vector machines (PSVMs)

Anirban Basudhar, Samy Missoum

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

This article presents a new probability of failure measure based on the notion of probabilistic support vector machines (PSVMs). A PSVM allows one to quantify the probability of having an error in the approximation of the failure boundary using a support vector machine (SVM). SVM can define explicitly the boundaries of disjoint and non-convex failure domains. The approximation of the failure boundary can be refined using an adaptive sampling scheme with a limited number of samples. However, the calculation of the probability of failure might still be inaccurate despite the adaptive sampling. In order to refine the probability estimate, the "quality" of the approximated boundary is quantified through the probability of misclassification of a sample by the SVM. A new measure of probability is then calculated using Monte-Carlo simulations that include the probability of misclassification. The proposed measure of probability of failure is such that it is always larger (i.e., more conservative) than the one obtained using a deterministic SVM. Several analytical examples are presented, including a case with two failure modes.

Original languageEnglish (US)
Title of host publicationCollection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference
StatePublished - 2010
Event51st AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference - Orlando, FL, United States
Duration: Apr 12 2010Apr 15 2010

Other

Other51st AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference
CountryUnited States
CityOrlando, FL
Period4/12/104/15/10

Fingerprint

Support vector machines
Sampling
Failure modes

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Mechanics of Materials
  • Building and Construction
  • Architecture

Cite this

Basudhar, A., & Missoum, S. (2010). Reliability assessment using probabilistic support vector machines (PSVMs). In Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference [2010-2764]

Reliability assessment using probabilistic support vector machines (PSVMs). / Basudhar, Anirban; Missoum, Samy.

Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference. 2010. 2010-2764.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Basudhar, A & Missoum, S 2010, Reliability assessment using probabilistic support vector machines (PSVMs). in Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference., 2010-2764, 51st AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Orlando, FL, United States, 4/12/10.
Basudhar A, Missoum S. Reliability assessment using probabilistic support vector machines (PSVMs). In Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference. 2010. 2010-2764
Basudhar, Anirban ; Missoum, Samy. / Reliability assessment using probabilistic support vector machines (PSVMs). Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference. 2010.
@inproceedings{bb2bb37ec653471fbf1bef69218f5755,
title = "Reliability assessment using probabilistic support vector machines (PSVMs)",
abstract = "This article presents a new probability of failure measure based on the notion of probabilistic support vector machines (PSVMs). A PSVM allows one to quantify the probability of having an error in the approximation of the failure boundary using a support vector machine (SVM). SVM can define explicitly the boundaries of disjoint and non-convex failure domains. The approximation of the failure boundary can be refined using an adaptive sampling scheme with a limited number of samples. However, the calculation of the probability of failure might still be inaccurate despite the adaptive sampling. In order to refine the probability estimate, the {"}quality{"} of the approximated boundary is quantified through the probability of misclassification of a sample by the SVM. A new measure of probability is then calculated using Monte-Carlo simulations that include the probability of misclassification. The proposed measure of probability of failure is such that it is always larger (i.e., more conservative) than the one obtained using a deterministic SVM. Several analytical examples are presented, including a case with two failure modes.",
author = "Anirban Basudhar and Samy Missoum",
year = "2010",
language = "English (US)",
isbn = "9781600867422",
booktitle = "Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference",

}

TY - GEN

T1 - Reliability assessment using probabilistic support vector machines (PSVMs)

AU - Basudhar, Anirban

AU - Missoum, Samy

PY - 2010

Y1 - 2010

N2 - This article presents a new probability of failure measure based on the notion of probabilistic support vector machines (PSVMs). A PSVM allows one to quantify the probability of having an error in the approximation of the failure boundary using a support vector machine (SVM). SVM can define explicitly the boundaries of disjoint and non-convex failure domains. The approximation of the failure boundary can be refined using an adaptive sampling scheme with a limited number of samples. However, the calculation of the probability of failure might still be inaccurate despite the adaptive sampling. In order to refine the probability estimate, the "quality" of the approximated boundary is quantified through the probability of misclassification of a sample by the SVM. A new measure of probability is then calculated using Monte-Carlo simulations that include the probability of misclassification. The proposed measure of probability of failure is such that it is always larger (i.e., more conservative) than the one obtained using a deterministic SVM. Several analytical examples are presented, including a case with two failure modes.

AB - This article presents a new probability of failure measure based on the notion of probabilistic support vector machines (PSVMs). A PSVM allows one to quantify the probability of having an error in the approximation of the failure boundary using a support vector machine (SVM). SVM can define explicitly the boundaries of disjoint and non-convex failure domains. The approximation of the failure boundary can be refined using an adaptive sampling scheme with a limited number of samples. However, the calculation of the probability of failure might still be inaccurate despite the adaptive sampling. In order to refine the probability estimate, the "quality" of the approximated boundary is quantified through the probability of misclassification of a sample by the SVM. A new measure of probability is then calculated using Monte-Carlo simulations that include the probability of misclassification. The proposed measure of probability of failure is such that it is always larger (i.e., more conservative) than the one obtained using a deterministic SVM. Several analytical examples are presented, including a case with two failure modes.

UR - http://www.scopus.com/inward/record.url?scp=84855617276&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84855617276&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781600867422

BT - Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference

ER -