Author | Kashif, Khan Falahudin |
Call Number | AIT Thesis no.ST-00-42 |
Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of
Engineering |
Publisher | Asian Institute of Technology |
Abstract | Econometric techniques are often used for modeling and forecasting construction demand.
Accuracy, as the vital attribute of assessing predictive abilities of all such models, has been the
topic of much research done in the developed countries. Not much research work has been
done in developing countries. This research thesis evaluates the accuracy of the three models
developed by a master degree student for modeling and forecasting sectoral construction
demand of Thailand. An aggregate demand model for Pakistan's construction industry is also
the part of this research.
The results of the analysis showed that the sectoral models of Thailand did not attain an
acceptable level of accuracy in either ex post forecasting or ex ante forecasting periods. The
results of fitting enlarged models to the relatively long series of data established that the
functional forms and inclusion of the same variables, used in the original models, might not
adequately improve the accuracy of the models. The implication of these findings was to
develop new models by including new variables and applying different functional forms for
improving the accuracy of the models. It was established by the results that multiple linear
regression techniques are suitable to model residential and other construction demand while
demand for non-residential construction is best represented by multiple log-linear regression
techniques. The total average accuracy gained by the new models over the original models, in
the ex post forecast period, was 61.3 percent.
Relative price index and the annual developmental expenditure mostly govern the demand for
Pakistan construction. The two variables could only represent the 56 percent of the variation
of the aggregate construction demand. The results of fitting seven functional forms to the
economic indicators of Pakistan showed that an important variable or variables, other than the
sixteen considered, might have improved the predictive performance of the models. The model
selected from the seven, attained accuracy of 12.24 in terms of mean absolute percentage
error. The general conclusion is that the proposed model can be used for ex ante forecasting in
the absence of any other modeling tool. |
Year | 2000 |
Type | Thesis |
School | School of Engineering and Technology (SET) |
Department | Department of Civil and Infrastucture Engineering (DCIE) |
Academic Program/FoS | Structural Engineering (STE) /Former Name = Structural Engineering and Construction (ST) |
Chairperson(s) | Ogunlana, Stephen 0.; |
Examination Committee(s) | Tang, John C. S.;Watanabe, Tsunemi ; |
Scholarship Donor(s) | Government of Japan; |
Degree | Thesis (M.Eng.) - Asian Institute of Technology,2000 |