Parameter Estimation for Scientists and Engineers

Adriaan Van Den Bos

• 07 augustus 2007
• 9780470147818

Samenvatting:

Parameter Estimation for Scientists and Engineers discusses estimating parameters of expectation models of statistical observations. It aims to show scientists and engineers, who often are not aware of estimators other than least squares, that statistical parameter estimation has much more to offer than least squares estimation alone.

Step-by-Step Methodology for Practical Parameter Estimation

Written for applied scientists and engineers, this book covers the most important aspects of estimating parameters of expectation models of statistical observations. The author demonstrates that statistical parameter estimation has much more to offer than least squares estimation alone, and explains how a priori knowledge may be used more fully to improve the precision of estimating. Parameter Estimation for Scientists and Engineers presents:

• An explanation of statistical parametric models of observations, and why they are used

• A description of distributions of observations including exponential families of distributions

• Fisher information and the Cramér-Rao lower bound, and how they are used to judge the quality of parameter estimators and experimental designs

• The maximum likelihood method and the least squares method for estimating parameters of expectation models

• A discussion of model hypothesis testing

• Numerical methods suitable for the parameter estimation problems dealt with in this book, as well as an exploration of how to use these methods in practice

Complete with sixty-two examples, eighty-nine problems and solutions, and thirty-four figures, Parameter Estimation for Scientists and Engineers is an invaluable reference for professionals and an ideal text for advanced undergraduate and graduate-level students in all disciplines of engineering and applied science.

The subject of this book is estimating parameters of expectation models of statistical observations. The book describes the most important aspects of the subject for applied scientists and engineers. This group of users is often not aware of estimators other than least squares. Therefore one purpose of this book is to show that statistical parameter estimation has much more to offer than least squares estimation alone. In the approach of this book, knowledge of the distribution of the observations is involved in the choice of estimators. A further advantage of the chosen approach is that it unifies the underlying theory and reduces it to a relatively small collection of coherent, generally applicable principles and notions.