/

Nematrian Reference Library

[this page | back links]

Set out below is information (held by the Nematrian website) on the reference you have selected


ReferenceTitleLink
Pankratz, A. (1991)Forecasting with Dynamic Regression Models

Paid link. As an Amazon Associate, Nematrian may earn commission from qualifying purchases made by you via this link.
Book cover (partial)

"Single equation regression models are one of the most widely used statistical forecasting tools in use today. Over the last two decades many ideas relevant to regression forecasting have arisen in the time series literature. For the professional statistician intent on a better understanding of data, better models, and better forecasts, a fundamental and detailed knowledge of regression methods is essential.

Designed as a companion volume to the author’s successful Forecasting with Univariate Box-Jenkins Models, this current text continues a discussion of Box-Jenkins modelling strategy, pulling together recent time series ideas in the Box-Jenkins tradition that are key to an informed practice of single equation forecasting.The text pays special attention to possible intertemporal (dynamic) patters – distributed lag responses of the output series (dependent variable) to the input series (independent variables), and the autocorrelation patterns of the regression disturbance.

Using step-by-step examples and several case studies, drawn largely from economics and business, the book clearly details, unlike many existing texts, just how elementary theory is directly applied to real data. The result is a unique hands-on seminar which includes:

- Cases with multiple inputs, both stochastic and deterministic

- Emphasis on model interpretation

- Emphasis on a model identification method that is easily applied with multiple inputs

- Practical rules for identifying rational form transfer functions

- Emphasis on feedback checks as a preliminary modelling step

- Development of an outlier and intervention detection method, including a BASIC computing routine

- A chapter linking dynamic regression models to vector ARMA models

- Use of the extended autorcorrelation function to identify ARIMA models"


See here to choose a new Category/Sub-Category or here for a list of all references held by the Nematrian website. Please contact us if any of the above material is inaccurate or if there are references you think should be included that we have excluded or vice-versa.
Desktop view | Switch to Mobile