Analysis of #house starts# for NZ Wool Board

By: Fradkin, L. (DSIR, Physics and Engineering Laboratory. Lower Hutt).
Contributor(s): DSIR, Physics and Engineering Laboratory. Lower Hutt.
Material type: materialTypeLabelBookSeries: Report / Physics and Engineering Laboratory ; 712.Publisher: 1981Description: 24 p.Report number: PEL-R--712Subject(s): MATHEMATICAL MODELS | PREDICTION | TIME SERIES ANALYSIS | LEAST SQUARES ANALYSIS | NEW ZEALAND WOOL BOARD
Incomplete contents:
The time-series A in this report was analysed for our client, the NZ Wool Board. Time series A consists of 69 successive <house starts>, the numbers of houses started to be built in New Zealand each quarter-year. The client is interested in identifying a mathematical model of this series to forecast the number of house starts one year ahead. In order to achieve this aim he attempted to apply the Box & Jenkins model building techniques to identify an auto-regressive model (i.e., a model in which the present value of the series is determined by its past values and a random shock). Our task was specified as a possible improvement of the resultant equation. In this report we go through all three stages of system identification: model structure identification, model identification and model validation. The first stage involves finding an appropriate type of parametric equation for description of the process. The second stage reduces to parameter estimation. At the third stage adequacy of the identified model is assessed for forecasting one year ahead. All three stages are very much interdependent
Tags from this library: No tags from this library for this title. Log in to add tags.
Item type Current location Call number Copy number Status Date due Barcode
ABSTRACT ABSTRACT NIWA BIBLIOGRAPHY
1 Available 31613-1001

4 refs; 4 figs; 4 tables

The time-series A in this report was analysed for our client, the NZ Wool Board. Time series A consists of 69 successive <house starts>, the numbers of houses started to be built in New Zealand each quarter-year. The client is interested in identifying a mathematical model of this series to forecast the number of house starts one year ahead. In order to achieve this aim he attempted to apply the Box & Jenkins model building techniques to identify an auto-regressive model (i.e., a model in which the present value of the series is determined by its past values and a random shock). Our task was specified as a possible improvement of the resultant equation. In this report we go through all three stages of system identification: model structure identification, model identification and model validation. The first stage involves finding an appropriate type of parametric equation for description of the process. The second stage reduces to parameter estimation. At the third stage adequacy of the identified model is assessed for forecasting one year ahead. All three stages are very much interdependent

PE

There are no comments on this title.

to post a comment.

Powered by Koha