Abstract
Water is of the essential needs for growth and global development. Bigger
numbers of population and industrialization leads to greater usage of water. As we
march to the area of modernization, actions must be considered to maintain the quality
and quality of water for the use of future generation. This project is intended to
forecast the limit of water consumption in the state of Perlis by using Box-Jenkins and
Neural Network Model. The main purpose of doing this research project is to help the
management in order to ensure that water supply will not be interrupted. The
interruption is predicted to be one of the major effects during drought where it is
commonly known that during this period residents will be supplied by only least amount
of water throughout the area. All the information pertaining to the percentage of water
usage (1997-2002) can be derived from Perlis Public Works Department,Water Supply
unit, whereas the percentage o f rain (1997-2002) will be available from Meteorological
Service department Mata Ayer,perlis. In order to decide which of two models is
preferable to determine the percentage of water usage in Perlis, results from both
method will be considered and compared. STATISTICA 6.0 and NeuralWorks
profesional II Plus are used for developing models. From the comparison we find that
result given by Box-Jenkins is better than Neural Network. Autoregressive(7) and
Autoregressive(9), which is respectively represents model for water usage and rain
to make prediction for two years ahead.
Website : http://ir.fsksm.utm.my/1419/
No comments:
Post a Comment