Wavelet decomposition-NNARX model for flood prediction of Kelantan River, Malaysia
Flood is a major disaster that happens around the world. It has caused the loss of many precious lives and massive destruction of property. The possibility of flood can be determined depends on many factors that consist of rainfall, structure of the river, flow rate of the river et...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/84219/1/FK%202019%2095%20-%20ir.pdf http://psasir.upm.edu.my/id/eprint/84219/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Flood is a major disaster that happens around the world. It has caused the loss of many precious
lives and massive destruction of property. The possibility of flood can be determined
depends on many factors that consist of rainfall, structure of the river, flow rate of the river
etc. One of the research challenges is to develop accurate prediction models and what improvement
can be made to the forecasting model. The objective of this thesis is to improve the performance of
the neural network model to predict the flood on the Kelantan River, Malaysia. A technique for
modelling of nonlinear data of flood forecasting using wavelet decomposition-neural network
autoregressive exogenous input (NNARX) approach is proposed.
This thesis discusses the identification of parameters that involved in the forecasting field as
rainfall value, flow rate of the river and the river water level. With the original data acquired,
the data had been processing through to wavelet decomposition and filtered to generate a new set of
input data for NNARX prediction model. This proposed technique has been compared with the
non-wavelet NNARX.
The experimental result show that the proposed approach provides better testing
performance compared to its counterpart, which the mean square error obtained is
2.0491e⁻⁴ while the normal NNARX is 6.1642e⁻⁴. |
|---|