Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting
The advancement of machine learning model has widely been adopted to provide flood forecast. However, the model must deal with the challenges to determine the most important features to be used in in flood forecast with high-dimensional non-linear time series when involving data from various station...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Universitas Ahmad Dahlan
2024
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/108221/ |
| _version_ | 1848865105088872448 |
|---|---|
| author | Mohd Khairudin, Nazli Mustapha, Norwati Mohd Aris, Teh Noranis Zolkepli, Maslina |
| author_facet | Mohd Khairudin, Nazli Mustapha, Norwati Mohd Aris, Teh Noranis Zolkepli, Maslina |
| author_sort | Mohd Khairudin, Nazli |
| building | UPM Institutional Repository |
| collection | Online Access |
| description | The advancement of machine learning model has widely been adopted to provide flood forecast. However, the model must deal with the challenges to determine the most important features to be used in in flood forecast with high-dimensional non-linear time series when involving data from various stations. Decomposition of time-series data such as empirical mode decomposition, ensemble empirical mode decomposition and discrete wavelet transform are widely used for optimization of input; however, they have been done for single dimension time-series data which are unable to determine relationships between data in high dimensional time series. In this study, hybrid machine learning models are developed based on this feature decomposition to forecast the monthly water level using monthly rainfall data. Rainfall data from eight stations in Kelantan River Basin are used in the hybrid model. To effectively select the best rainfall data from the multi-stations that provide higher accuracy, these rainfall data are analyzed with entropy called Mutual Information that measure the uncertainty of random variables from various stations. Mutual Information act as optimization method helps the researcher to select the appropriate features to score higher accuracy of the model. The experimental evaluations proved that the hybrid machine learning model based on the feature decomposition and ranked by Mutual Information can increase the accuracy of water level forecasting. This outcome will help the authorities in managing the risk of flood and helping people in the evacuation process as an early warning can be assigned and disseminate to the citizen. |
| first_indexed | 2025-11-15T13:59:25Z |
| format | Article |
| id | upm-108221 |
| institution | Universiti Putra Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-15T13:59:25Z |
| publishDate | 2024 |
| publisher | Universitas Ahmad Dahlan |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | upm-1082212024-09-11T03:34:11Z http://psasir.upm.edu.my/id/eprint/108221/ Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting Mohd Khairudin, Nazli Mustapha, Norwati Mohd Aris, Teh Noranis Zolkepli, Maslina The advancement of machine learning model has widely been adopted to provide flood forecast. However, the model must deal with the challenges to determine the most important features to be used in in flood forecast with high-dimensional non-linear time series when involving data from various stations. Decomposition of time-series data such as empirical mode decomposition, ensemble empirical mode decomposition and discrete wavelet transform are widely used for optimization of input; however, they have been done for single dimension time-series data which are unable to determine relationships between data in high dimensional time series. In this study, hybrid machine learning models are developed based on this feature decomposition to forecast the monthly water level using monthly rainfall data. Rainfall data from eight stations in Kelantan River Basin are used in the hybrid model. To effectively select the best rainfall data from the multi-stations that provide higher accuracy, these rainfall data are analyzed with entropy called Mutual Information that measure the uncertainty of random variables from various stations. Mutual Information act as optimization method helps the researcher to select the appropriate features to score higher accuracy of the model. The experimental evaluations proved that the hybrid machine learning model based on the feature decomposition and ranked by Mutual Information can increase the accuracy of water level forecasting. This outcome will help the authorities in managing the risk of flood and helping people in the evacuation process as an early warning can be assigned and disseminate to the citizen. Universitas Ahmad Dahlan 2024 Article PeerReviewed Mohd Khairudin, Nazli and Mustapha, Norwati and Mohd Aris, Teh Noranis and Zolkepli, Maslina (2024) Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting. International Journal of Advances in Intelligent Informatics, 10 (1). pp. 1-13. ISSN 2442-6571; ESSN: 2548-3161 http://ijain.org/index.php/IJAIN/article/view/1130 10.26555/ijain.v10i1.1130 |
| spellingShingle | Mohd Khairudin, Nazli Mustapha, Norwati Mohd Aris, Teh Noranis Zolkepli, Maslina Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| title | Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| title_full | Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| title_fullStr | Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| title_full_unstemmed | Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| title_short | Hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| title_sort | hybrid machine learning model based on feature decomposition and entropy optimization for higher accuracy flood forecasting |
| url | http://psasir.upm.edu.my/id/eprint/108221/ http://psasir.upm.edu.my/id/eprint/108221/ http://psasir.upm.edu.my/id/eprint/108221/ |