dc.contributor.author Nikzad-Langerodi, R. dc.contributor.author Lughofer, E. dc.contributor.author Cernuda, C. dc.contributor.author Reischer, T. dc.contributor.author Kantner, W. dc.contributor.author Pawliczek, M. dc.contributor.author Brandstetter, M. dc.date 2018-11-08 en_US dc.date.accessioned 2018-11-28T14:28:14Z dc.date.available 2018-11-28T14:28:14Z dc.date.issued 2018 dc.identifier.issn 0003-2670 dc.identifier.uri http://hdl.handle.net/20.500.11824/891 dc.description.abstract The physico-chemical properties of Melamine Formaldehyde (MF) based thermosets are largely influenced by the degree of polymerization (DP) in the underlying resin. On-line supervision of the turbidity point by means of vibrational spectroscopy has recently emerged as a promising technique to monitor the DP of MF resins. However, spectroscopic determination of the DP relies on chemometric models, which are usually sensitive to drifts caused by instrumental and/or sample associated changes occurring over time. In order to detect the time point when drifts start causing prediction bias, we here explore a universal drift detector based on a faded version of the Page-Hinkley (PH) statistic, which we test in three data streams from an industrial MF resin production process. We employ committee disagreement (CD), computed as the variance of model predictions from an ensemble of partial least squares (PLS) models, as a measure for sample-wise prediction uncertainty and use the PH statistic to detect hanges in this quantity. We further explore supervised and unsupervised strategies for (semi-)automatic model adaptation upon detection of a drift. For the former, manual reference measurements are requested whenever statistical thresholds on Hotelling’s $T^2$ and/or Q-Residuals are violated. Models are subsequently re-calibrated using weighted partial least squares in order to increase the influence of newer samples, which increases the flexibility when adapting to new (drifted) states. Unsupervised model adaptation is carried out exploiting the dual antecedent-consequent structure of a recently developed fuzzy systems variant of PLS termed FLEXFIS-PLS. In particular, antecedent parts are updated while maintaining the internal structure of the local linear predictors (i.e. the consequents). We found improved drift detection capability of the CD compared to Hotelling’s $T^2$ and Q-Residuals when used in combination with the proposed PH test. Furthermore, we found that active selection of samples by active learning (AL) used for subsequent model adaptation is advantageous compared to passive (random) selection in case that a drift leads to persistent prediction bias allowing more rapid adaptation at lower reference measurement rates. Fully unsupervised adaptation using FLEXFIS-PLS could improve predictive accuracy significantly for light drifts but was not able to fully compensate for prediction bias in case of significant lack of fit w.r.t. the latent variable space. en_US dc.format application/pdf en_US dc.language.iso eng en_US dc.rights Reconocimiento-NoComercial-CompartirIgual 3.0 España en_US dc.rights.uri http://creativecommons.org/licenses/by-nc-sa/3.0/es/ en_US dc.subject ensembled PLS, drift detection, committee disagreement, active learning, calibration model adaptation, weighted learning, Melamine formaldehyde resin en_US dc.title Calibration Model Maintenance in Melamine Resin Production: Integrating Drift Detection, Smart Sample Selection and Model Adaptation en_US dc.type info:eu-repo/semantics/article en_US dc.identifier.doi 10.1016/j.aca.2018.02.003 dc.relation.publisherversion https://www.sciencedirect.com/science/article/pii/S0003267018301752 en_US dc.relation.projectID ES/1PE/SEV-2013-0323 en_US dc.relation.projectID ES/1PE/TIN2017-82626-R en_US dc.relation.projectID EUS/BERC/BERC.2014-2017 en_US dc.rights.accessRights info:eu-repo/semantics/embargoedAccess en_US dc.type.hasVersion info:eu-repo/semantics/acceptedVersion en_US dc.journal.title Analytica Chimica Acta en_US
﻿

### This item appears in the following Collection(s)

Except where otherwise noted, this item's license is described as Reconocimiento-NoComercial-CompartirIgual 3.0 España