Almost autonomous training of mixtures of principal component analyzers

Musa, Mohamed E. M.; Ridder, Dick de; Duin, Robert P. W.; Atalay, Volkan

doi:10.1016/j.patrec.2004.03.019

Cited by 6 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Exhaustive search over all possible models is inefficient due to the large number of candidate models to be considered. A fast sub-optimal algorithm with Bayesian information criterion (BIC) [14,26] was proposed to address the model selection problem.…”

Section: Propositionmentioning

confidence: 99%

Reconstruction-based multivariate contribution analysis for fault isolation: A branch and bound approach

Yang

Chen

et al. 2012

Journal of Process Control

View full text Add to dashboard Cite

-Identification of faulty variables is an important component of multivariate statistical process monitoring (MSPM); it provides crucial information for further analysis of the root cause of the detected fault. The main challenge is the large number of combinations of process variables under consideration, usually resulting in a combinatorial optimization problem. This paper develops a generic reconstruction based multivariate contribution analysis (RBMCA) framework to identify the variables that are the most responsible for the fault. A branch and bound (BAB) algorithm is proposed to efficiently solve the combinatorial optimization problem. The formulation of the RBMCA does not depend on a specific model, which allows it to be applicable to any MSPM model. We demonstrate the application of the RBMCA to a specific model: the mixture of probabilistic principal component analysis (PPCA mixture) model. Finally, we illustrate the effectiveness and computational efficiency of the proposed methodology through a numerical example and the benchmark simulation of the Tennessee Eastman process.

show abstract

Section: Propositionmentioning

confidence: 99%

Reconstruction-based multivariate contribution analysis for fault isolation: A branch and bound approach

Yang

Chen

et al. 2012

Journal of Process Control

View full text Add to dashboard Cite

show abstract

“…According to the criterion of variance ratio, the retained PCs should explain in excess of say, 90%, of the total variance in the data (Musa et al, 2004).…”

Section: Model Selectionmentioning

confidence: 99%

“…In a fast model selection algorithm was proposed with the restrictive assumption that all mixtures 6 have the same number of PCs (Q k ). To allow different Q k 's while attaining computational efficiency, Musa et al (2004) proposed to apply the criterion of variance ratio for the selection of PCs within each mixture, and developed a greedy training algorithm to determine the number of mixtures. Despite the fact that this two-stage algorithm is sub-optimal, it was shown (Musa et al, 2004) to achieve promising performance in practice.…”

Section: Model Selectionmentioning

confidence: 99%