A hierarchical framework for simultaneous analysis of multiple related individual datasets is pre... more A hierarchical framework for simultaneous analysis of multiple related individual datasets is presented. The approach is very similar to mixed effects modelling as known from statistical theory. The model used at the individual level is, in principle, irrelevant as long as a maximum likelihood estimate and its uncertainty (Hessian) can be computed. The individual model used in this text is a hidden Markov model. A simulation study concerning a two-dimensional biased random walk is examined to verify the consistency of the hierarchical estimation framework. In addition, a study based on acoustic telemetry data from pike illustrates how the framework can identify individuals that deviate from the remaining population.
The real-time fault monitoring and control of the Vacuum Assisted Resin Transfer Moulding (VARTM)... more The real-time fault monitoring and control of the Vacuum Assisted Resin Transfer Moulding (VARTM) production process requires a knowledge of the position of the epoxy flow-front inside the mould. Therefore, a fast and accurate flow-front tracking system capable of combining the underlying physics of the flow-front dynamics with the measured data is highly prized. Stochastic differential equations (SDEs) based grey-box models deliver a good trade-off between high fidelity models and data-driven black-box models for designing such a flow-front position tracking system. In this paper, we propose a simple yet novel coupled SDE based spatiotemporal grey-box model of the flow-front dynamics in case of missing sensor information. The proposed method uses the finite difference approximation of the spatial domain of the flow-front for estimating spatial flow pattern of the epoxy. Furthermore, to accommodate for the missing sensor data, we utilise a modified version of the continuous-discrete...
... DK-2920 Charlottenlund, Denmark. 2University of Hawaii at Manoa, Hawaii 96822 USA 3Leibniz-In... more ... DK-2920 Charlottenlund, Denmark. 2University of Hawaii at Manoa, Hawaii 96822 USA 3Leibniz-Institute of Marine Research, Düsternbrooker Weg 20, D 24105 Kiel, Germany August 23, 2006 Abstract A central ingredient ...
6 We present a statistical method for intercalibration of surveys, i.e. determining 7 the relativ... more 6 We present a statistical method for intercalibration of surveys, i.e. determining 7 the relative selectivity of two gear types or two vessels, based on data from paired 8 trawl hauls. The model estimates not only the relative selectivity in each length class, 9 but also the size spectrum of the underlying population at each station, and takes 10 into account size-structured clustering of fish at small temporal and spatial scales. 11 We assume Poisson distributed catches conditional on log-Gaussian variables, which 12 allows for overdispersion and correlation between catch counts in neighboring size 13 classes. We apply the method to catches of hake (M. Paradoxus and M. Capensis) 14 in paired trawl hauls performed by two different vessels, viz. the RV Dr. Fridtjof 15 Nansen and the FV Blue Sea. The results demonstrate that it is feasible to estimate 16 the relative selectivity in each size class, and to test statistically the hypothesis that 17 the selectivity is independent of siz...
In contrast to previous approaches, and incorporating interactions between individual prey in the... more In contrast to previous approaches, and incorporating interactions between individual prey in the stomach, the cylindermodel of gastric evacuation predicted accurately the ingestion times of individual prey recovered from stomachs of predatory gadoids sampled in laboratory experiments. For application to field situations, estimates of the variance σ2 τ = (σ2 τ,e + σ2 τ,m) of the predicted time interval τ between prey ingestion and stomach samplingwere obtained from generalised considerations about the errors of estimated gastric evacuation rate (στ,e ∼= 0.10τ) and prey size at ingestion (στ,m increased from 1 h to 2.5 h with increasing number and body size of other prey in the stomach). The bias originating from non-observable prey (that were completely evacuated from the stomach prior to sampling) may amount to more than 20% and should be taken into consideration. In contrast, the sensitivity of τ to estimated body lengths of other prey recovered as small remains from the stomach w...
A hierarchical framework for simultaneous analysis of multiple related individual datasets is pre... more A hierarchical framework for simultaneous analysis of multiple related individual datasets is presented. The approach is very similar to mixed effects modelling as known from statistical theory. The model used at the individual level is, in principle, irrelevant as long as a maximum likelihood estimate and its uncertainty (Hessian) can be computed. The individual model used in this text is a hidden Markov model. A simulation study concerning a two-dimensional biased random walk is examined to verify the consistency of the hierarchical estimation framework. In addition, a study based on acoustic telemetry data from pike illustrates how the framework can identify individuals that deviate from the remaining population.
The real-time fault monitoring and control of the Vacuum Assisted Resin Transfer Moulding (VARTM)... more The real-time fault monitoring and control of the Vacuum Assisted Resin Transfer Moulding (VARTM) production process requires a knowledge of the position of the epoxy flow-front inside the mould. Therefore, a fast and accurate flow-front tracking system capable of combining the underlying physics of the flow-front dynamics with the measured data is highly prized. Stochastic differential equations (SDEs) based grey-box models deliver a good trade-off between high fidelity models and data-driven black-box models for designing such a flow-front position tracking system. In this paper, we propose a simple yet novel coupled SDE based spatiotemporal grey-box model of the flow-front dynamics in case of missing sensor information. The proposed method uses the finite difference approximation of the spatial domain of the flow-front for estimating spatial flow pattern of the epoxy. Furthermore, to accommodate for the missing sensor data, we utilise a modified version of the continuous-discrete...
... DK-2920 Charlottenlund, Denmark. 2University of Hawaii at Manoa, Hawaii 96822 USA 3Leibniz-In... more ... DK-2920 Charlottenlund, Denmark. 2University of Hawaii at Manoa, Hawaii 96822 USA 3Leibniz-Institute of Marine Research, Düsternbrooker Weg 20, D 24105 Kiel, Germany August 23, 2006 Abstract A central ingredient ...
6 We present a statistical method for intercalibration of surveys, i.e. determining 7 the relativ... more 6 We present a statistical method for intercalibration of surveys, i.e. determining 7 the relative selectivity of two gear types or two vessels, based on data from paired 8 trawl hauls. The model estimates not only the relative selectivity in each length class, 9 but also the size spectrum of the underlying population at each station, and takes 10 into account size-structured clustering of fish at small temporal and spatial scales. 11 We assume Poisson distributed catches conditional on log-Gaussian variables, which 12 allows for overdispersion and correlation between catch counts in neighboring size 13 classes. We apply the method to catches of hake (M. Paradoxus and M. Capensis) 14 in paired trawl hauls performed by two different vessels, viz. the RV Dr. Fridtjof 15 Nansen and the FV Blue Sea. The results demonstrate that it is feasible to estimate 16 the relative selectivity in each size class, and to test statistically the hypothesis that 17 the selectivity is independent of siz...
In contrast to previous approaches, and incorporating interactions between individual prey in the... more In contrast to previous approaches, and incorporating interactions between individual prey in the stomach, the cylindermodel of gastric evacuation predicted accurately the ingestion times of individual prey recovered from stomachs of predatory gadoids sampled in laboratory experiments. For application to field situations, estimates of the variance σ2 τ = (σ2 τ,e + σ2 τ,m) of the predicted time interval τ between prey ingestion and stomach samplingwere obtained from generalised considerations about the errors of estimated gastric evacuation rate (στ,e ∼= 0.10τ) and prey size at ingestion (στ,m increased from 1 h to 2.5 h with increasing number and body size of other prey in the stomach). The bias originating from non-observable prey (that were completely evacuated from the stomach prior to sampling) may amount to more than 20% and should be taken into consideration. In contrast, the sensitivity of τ to estimated body lengths of other prey recovered as small remains from the stomach w...
Uploads
Papers