A Transductive Learning Approach to Process Fault Identification

The problem of fault identification is considered using recent developments in machine learning that allow the use of unlabeled data to optimally define decision boundaries that separate data belonging to different categories. Traditionally, fault identification uses either hardware redundancy or so...

Full description

Bibliographic Details
Main Authors: Jemwa, G., Aldrich, Chris
Other Authors: Gorden T. Jemwa
Format: Conference Paper
Published: Elsevier 2010
Subjects:
Online Access:http://hdl.handle.net/20.500.11937/15577
_version_ 1848748930633826304
author Jemwa, G.
Aldrich, Chris
author2 Gorden T. Jemwa
author_facet Gorden T. Jemwa
Jemwa, G.
Aldrich, Chris
author_sort Jemwa, G.
building Curtin Institutional Repository
collection Online Access
description The problem of fault identification is considered using recent developments in machine learning that allow the use of unlabeled data to optimally define decision boundaries that separate data belonging to different categories. Traditionally, fault identification uses either hardware redundancy or software redundancy for trouble-shooting the source of faults in a system. Unfortunately, this imposes a data redundancy cost on such systems. Instead of performing model inversion that requires an accurate model, in transduction estimates of the values of a function at specified points are required, instead of learning a general rule on the entire input domain. Transductive learning is motivated from similar arguments underlying state-of-the-art classification and regression methods such as support vector machines. However, transduction is more fundamental as it is a step used in proving learning error bounds in classical statistical learning theory. Use of transduction allows a flexible ordering of the classes of functions from which a model is selected and, therefore, the error bounds are provably tight. The potential of the proposed framework is assessed using data from metallurgical process systems. It is shown that for higher dimensional and large multiclass systems, the proposed framework gives betterperformances with respect to classification error minimization.
first_indexed 2025-11-14T07:12:52Z
format Conference Paper
id curtin-20.500.11937-15577
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T07:12:52Z
publishDate 2010
publisher Elsevier
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-155772023-01-13T07:56:29Z A Transductive Learning Approach to Process Fault Identification Jemwa, G. Aldrich, Chris Gorden T. Jemwa Chris Aldrich Fault Diagnosis Machine Learning Fault Identification Pattern Recognition The problem of fault identification is considered using recent developments in machine learning that allow the use of unlabeled data to optimally define decision boundaries that separate data belonging to different categories. Traditionally, fault identification uses either hardware redundancy or software redundancy for trouble-shooting the source of faults in a system. Unfortunately, this imposes a data redundancy cost on such systems. Instead of performing model inversion that requires an accurate model, in transduction estimates of the values of a function at specified points are required, instead of learning a general rule on the entire input domain. Transductive learning is motivated from similar arguments underlying state-of-the-art classification and regression methods such as support vector machines. However, transduction is more fundamental as it is a step used in proving learning error bounds in classical statistical learning theory. Use of transduction allows a flexible ordering of the classes of functions from which a model is selected and, therefore, the error bounds are provably tight. The potential of the proposed framework is assessed using data from metallurgical process systems. It is shown that for higher dimensional and large multiclass systems, the proposed framework gives betterperformances with respect to classification error minimization. 2010 Conference Paper http://hdl.handle.net/20.500.11937/15577 Elsevier restricted
spellingShingle Fault Diagnosis
Machine Learning
Fault Identification
Pattern Recognition
Jemwa, G.
Aldrich, Chris
A Transductive Learning Approach to Process Fault Identification
title A Transductive Learning Approach to Process Fault Identification
title_full A Transductive Learning Approach to Process Fault Identification
title_fullStr A Transductive Learning Approach to Process Fault Identification
title_full_unstemmed A Transductive Learning Approach to Process Fault Identification
title_short A Transductive Learning Approach to Process Fault Identification
title_sort transductive learning approach to process fault identification
topic Fault Diagnosis
Machine Learning
Fault Identification
Pattern Recognition
url http://hdl.handle.net/20.500.11937/15577