Data-driven modelling

Teegavarapu RSV, Chandramouli V (2005) Improved weighting methods, deterministic and stochastic data-driven models for estimation of missing precipitation records. J Hydrol 312 191-206... 

K. Yamuna Rani and S.C. Patwardhan. Data-driven model based control of a multi-product... 

How NUMA is handled by an automatic parallelizing compiler that hides all complexity, by a data-driven model in which distribution of data between processors is made explicit but all data are referenced with the same language constructs (e.g.. High Performance FORTRAN), or by a subroutine interface to access distributed or remote data (e.g., Linda tuples) or it may be that no direct access is provided to remote data (e.g., message passing). 

Janes KA, Yaffe MB. Data-driven modelling of signal-transduction networks. Nat. Rev. Mol. Cell. Biol. 2006 7 820-828. 

Archer D. E. (1996b) A data-driven model of the global calcite lysocline. Global Biogeochem. Cycles 10, 511-526. 

A wholistic scale-up strategy consists of a comprehensive and detailed process characterization to identify key stress factors and key parameters influencing product yield and quality the most, and of an appropriate process control and process design ensuring optimum mixing and reaction conditions, supported by appropriate knowledge and data-driven models [234,235] as well as computational tools. 

The building of data-driven models in this study was constrained by the sprarsity of the available data, as there were only 18 independent samples (rocket motor designs) available. In addition the input and output data were highly multivariate with 18 rocket motor design parameters and 146 spectral wavelengths in the middle IR band. One advantage is that the IR spectral measurements were repeated a number of times (4 to 44 repeats per rocket motor). 

First-principle modelling Data-driven modeling... 

Solomatine, D.R 2002. Data-driven modelling Paradigm, methods, experiences . Proc. 5th International Conference on Hyroinformatics, UK. 

To be an effective management tool, productivity management should be linked to a data-information decision-making model. Figure 16.7 proposes an eHealth data-driven model that identifies customer/patient value expectations (value proposition), the availability of resource requirements and their deployment, and the relationship between an acceptable (viable) value proposition and organizational profitability, productivity, and competitive advantage. 

Data-driven or black-box modelling, where a description of the process is obtained solely by developing models for the available data. This approach can provide very accurate models of the system at a given set of conditions, but the model cannot generalise well to other conditions. Furthermore, developing such models can be difficult, since the selection of appropriate terms and relevant data is a nontrivial task. Unless the correlations are strong, it may be difficult to decide on an appropriate data-driven model. 

Given the potential problems associated with both approaches, a third, middle way, has also been considered. This approach is called grey-box modelling, where the initial form of the equation determined based on the first-principle model is used for data-driven modelling. This approach has the advantage that the form of the equation has some physical meaning and could provide a reasonable description of the process. 

Data-driven models can be used for arbitrary conditions and operating points. 

Grey-box modelling combines the advantages of first-principle and data-driven models. 

Ventura, A. C., Bruno, L., and Dawson, S. P. (2006). Simple data-driven models of intracellular calcium dynamics with predictive power. Phys. Rev. E. Stat. Nonlin. Soft. Matter Phys. 74, 011917. 

In conclusion, the HtL phase involves the integration of multiple data sources to design and test hypotheses around chemical scope. As the pharmaceutical industry continues to evolve toward attrition-driven and data-driven models, this... 

This relationship can be either analytical expression or data-driven models in accordance with the system complexity. 

Conventional layer classification is the first step for product analysis. Then, the related failure mode and mechanism, environmental and load conditions, and other information should be determined to conduct performance simulation in component level. Through the transfer function between system and component levels, performance simulation in system level can be carried out, collecting simulated performance data varying with time under specific conditions. Finally, modelling the degradation process with consideration of the failure threshold is to produce the model set M,. Noting that theoretically physical model can be established for product with simple failure mechanism, while modem engineering tools should be used for complicated products. Thus, the model set M, can be analytical functions or data-driven models. 

---