| Summary: | 碩士 === 輔仁大學 === 應用統計學研究所 === 94 === In addition to represent the price for construction materials and labors, Construction Cost Index (CCI) provides with the function to adjust outlays for constructions in progress. Forecasting for price trend of construction materials and labors provides information not only for government organizations to evaluate reasonableness of outlays of public constructions, but also for small and medium construction firms to measure profit of constructions. Although CCI is a key indicator for construction cost and government organizations, the authoritative forecasting for CCI is deficient. To this end, the study integrate Gene Algorithm(GA) and invert model analysis which is in common use for security markets to develop methodology for forecasting CCI trend and reverse point by application of Microsoft Excel extra gene algorithm module-Evolver.
Due to lack original transaction data, the study adopts the price and weight data from January 2001 to October 2005 for 106 construction material and labor items proclaimed by Directorate-General of Budget, Accounting and Statistics. To develop the forecasting methodology, the study follows the several steps including diminishing the range of items from 106 items to 14 common use items, setting computing rules and limitation terms for gene algorithm, building index computing models, computing trend baseline of CCI, building interval of trend baseline of CCI.
The result, that error of CCI between 106 items and 14 items is within 4.83%, shows that the methodology is helpful to simplify collecting data and computing index. Without the difficulty of gene algorithm, application of Evolver is so easy to use that the cost of forecasting is very low and suit for small and medium construction firms. Besides, the methodology can be applied to fit forecasting needs in other fields by modifying price items and parameters.
|
|---|
