Dynamic Error-in-variables Models

Authors

  • Vincenzo G. Dovì ISTIC "G.B.Bonino" - Genoa University, Genova (Italy)

DOI:

https://doi.org/10.17713/ajs.v27i1&2.526

Abstract

Several environmental problems are described by models defined by systems of first order differential equations. Typically, the derivatives of concentrations are expressed as functions of independent variables such as temperature, pressure, pH, etc. However, the independent variables are generally known only through measurements made at discrete times. Thus we have to face two types of difficulties
1) estimate the errors in the measurements of the independent variables
2) assume suitable functional forms for them in the intervals between two
successive measurements.
In analogy to the algebraic case, we have chosen to define this class of problems as “dynamic error-in-variables” models. Furthermore we have developed an algorithm to carry out the general estimation problem. An example is described in detail for illustration purposes.

References

V.G. Dovì. Evaluating the Bias in Kinetic Parameters Estimation due to Temperature Fluctuations. To be published.

V.G. Dovì, L. Maga, and A.P. Reverberi. On the Identification of Approximate Models, Comput. Chem. Engng., 20, (11), 1361-1367, 1996.

V.G. Dovì, E. Arato, and L. Maga. A More General Formulation of Separable Least Squares, Mathematical Modelling 18, 20-23, 1985.

Y. Bard. Nonlinear Parameter Estimation, Academic Press, New York, 1974.

Downloads

Published

2016-04-03

Issue

Section

Articles

How to Cite

Dynamic Error-in-variables Models. (2016). Austrian Journal of Statistics, 27(1&2), 7-15. https://doi.org/10.17713/ajs.v27i1&2.526