Rainfall modeling using two different neural networks improved by metaheuristic algorithms
Abstract Rainfall is crucial for the development and management of water resources. Six hybrid soft computing models, including multilayer perceptron (MLP)–Henry gas solubility optimization (HGSO), MLP–bat algorithm (MLP–BA), MLP–particle swarm optimization (MLP–PSO), radial basis neural network fun...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-12-01
|
| Series: | Environmental Sciences Europe |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12302-023-00818-0 |
| _version_ | 1827581786640613376 |
|---|---|
| author | Saad Sh. Sammen Ozgur Kisi Mohammad Ehteram Ahmed El-Shafie Nadhir Al-Ansari Mohammad Ali Ghorbani Shakeel Ahmad Bhat Ali Najah Ahmed Shamsuddin Shahid |
| author_facet | Saad Sh. Sammen Ozgur Kisi Mohammad Ehteram Ahmed El-Shafie Nadhir Al-Ansari Mohammad Ali Ghorbani Shakeel Ahmad Bhat Ali Najah Ahmed Shamsuddin Shahid |
| author_sort | Saad Sh. Sammen |
| collection | DOAJ |
| description | Abstract Rainfall is crucial for the development and management of water resources. Six hybrid soft computing models, including multilayer perceptron (MLP)–Henry gas solubility optimization (HGSO), MLP–bat algorithm (MLP–BA), MLP–particle swarm optimization (MLP–PSO), radial basis neural network function (RBFNN)–HGSO, RBFNN–PSO, and RBFGNN–BA, were used in this study to forecast monthly rainfall at two stations in Malaysia (Sara and Banding). Different statistical measures (mean absolute error (MAE) and Nash–Sutcliffe efficiency (NSE) and percentage of BIAS (PBIAS)) and a Taylor diagram were used to assess the models’ performance. The results indicated that the MLP–HGSO performed better than the other models in forecasting rainfall at both stations. In addition, transition matrices were computed for each station and year based on the conditional probability of rainfall or absence of rainfall on a given month. The values of MAE for testing processes for the MLP–HGSO, MLP–PSO, MLP–BA, RBFNN–HGSO, RBFNN–BA, and RBFNN–PSO at the first station were 0.712, 0.755, 0.765, 0.717, 0.865, and 0.891, while the corresponding NSE and PBIAS values were 0.90–0.23, 0.83–0.29, 0.85–0.25, 0.87–0.27, 0.81–0.31, and 0.80–0.35, respectively. For the second station, the values of MAE were found 0.711, 0.743, 0.742, 0.719, 0.863 and 0.890 for the MLP–HGSO, MLP–PSO, MLP–BA, RBFNN–HGSO, RBFNN–BA, and RBFNN–PSO during testing processes and the corresponding NSE–PBIAS values were 0.92–0.22, 0.85–0.28, 0.89–0.26, 0.91–0.25, 0.83–0.31, 0.82–0.32, respectively. Based on the outputs of the MLP–HGSO, the highest rainfall was recorded in 2012 with a probability of 0.72, while the lowest rainfall was recorded in 2006 with a probability of 0.52 at the Sara Station. In addition, the results indicated that the MLP–HGSO performed better than the other models within the Banding Station. According to the findings, the hybrid MLP–HGSO was selected as an effective rainfall prediction model. |
| first_indexed | 2024-03-08T22:41:22Z |
| format | Article |
| id | doaj.art-6fcaa6591d5a4ea5abf4987408f0cc59 |
| institution | Directory Open Access Journal |
| issn | 2190-4715 |
| language | English |
| last_indexed | 2024-03-08T22:41:22Z |
| publishDate | 2023-12-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Environmental Sciences Europe |
| spelling | doaj.art-6fcaa6591d5a4ea5abf4987408f0cc592023-12-17T12:09:46ZengSpringerOpenEnvironmental Sciences Europe2190-47152023-12-0135111610.1186/s12302-023-00818-0Rainfall modeling using two different neural networks improved by metaheuristic algorithmsSaad Sh. Sammen0Ozgur Kisi1Mohammad Ehteram2Ahmed El-Shafie3Nadhir Al-Ansari4Mohammad Ali Ghorbani5Shakeel Ahmad Bhat6Ali Najah Ahmed7Shamsuddin Shahid8Department of Civil Engineering, College of Engineering, University of DiyalaDepartment of Civil Engineering, Technical University of LübeckDepartment of Water Engineering and Hydraulic Structures, Faculty of Civil EngineeringCivil Engineering Department, Faculty of Engineering, University of MalayaCivil, Environmental and Natural Resources Engineering, Lulea University of TechnologyDepartment of Civil Engineering, Istanbul Technical UniversityCollege of Agricultural Engineering and Technology, SKUAST-KashmirInstitute of Energy Infrastructure (IEI), Universiti Tenaga Nasional (UNITEN)School of Civil Engineering, Faculty of Engineering, Universiti Teknologi Malaysia (UTM)Abstract Rainfall is crucial for the development and management of water resources. Six hybrid soft computing models, including multilayer perceptron (MLP)–Henry gas solubility optimization (HGSO), MLP–bat algorithm (MLP–BA), MLP–particle swarm optimization (MLP–PSO), radial basis neural network function (RBFNN)–HGSO, RBFNN–PSO, and RBFGNN–BA, were used in this study to forecast monthly rainfall at two stations in Malaysia (Sara and Banding). Different statistical measures (mean absolute error (MAE) and Nash–Sutcliffe efficiency (NSE) and percentage of BIAS (PBIAS)) and a Taylor diagram were used to assess the models’ performance. The results indicated that the MLP–HGSO performed better than the other models in forecasting rainfall at both stations. In addition, transition matrices were computed for each station and year based on the conditional probability of rainfall or absence of rainfall on a given month. The values of MAE for testing processes for the MLP–HGSO, MLP–PSO, MLP–BA, RBFNN–HGSO, RBFNN–BA, and RBFNN–PSO at the first station were 0.712, 0.755, 0.765, 0.717, 0.865, and 0.891, while the corresponding NSE and PBIAS values were 0.90–0.23, 0.83–0.29, 0.85–0.25, 0.87–0.27, 0.81–0.31, and 0.80–0.35, respectively. For the second station, the values of MAE were found 0.711, 0.743, 0.742, 0.719, 0.863 and 0.890 for the MLP–HGSO, MLP–PSO, MLP–BA, RBFNN–HGSO, RBFNN–BA, and RBFNN–PSO during testing processes and the corresponding NSE–PBIAS values were 0.92–0.22, 0.85–0.28, 0.89–0.26, 0.91–0.25, 0.83–0.31, 0.82–0.32, respectively. Based on the outputs of the MLP–HGSO, the highest rainfall was recorded in 2012 with a probability of 0.72, while the lowest rainfall was recorded in 2006 with a probability of 0.52 at the Sara Station. In addition, the results indicated that the MLP–HGSO performed better than the other models within the Banding Station. According to the findings, the hybrid MLP–HGSO was selected as an effective rainfall prediction model.https://doi.org/10.1186/s12302-023-00818-0MLPRBFNNProbability matrixMarkov chainRainfall modelling |
| spellingShingle | Saad Sh. Sammen Ozgur Kisi Mohammad Ehteram Ahmed El-Shafie Nadhir Al-Ansari Mohammad Ali Ghorbani Shakeel Ahmad Bhat Ali Najah Ahmed Shamsuddin Shahid Rainfall modeling using two different neural networks improved by metaheuristic algorithms Environmental Sciences Europe MLP RBFNN Probability matrix Markov chain Rainfall modelling |
| title | Rainfall modeling using two different neural networks improved by metaheuristic algorithms |
| title_full | Rainfall modeling using two different neural networks improved by metaheuristic algorithms |
| title_fullStr | Rainfall modeling using two different neural networks improved by metaheuristic algorithms |
| title_full_unstemmed | Rainfall modeling using two different neural networks improved by metaheuristic algorithms |
| title_short | Rainfall modeling using two different neural networks improved by metaheuristic algorithms |
| title_sort | rainfall modeling using two different neural networks improved by metaheuristic algorithms |
| topic | MLP RBFNN Probability matrix Markov chain Rainfall modelling |
| url | https://doi.org/10.1186/s12302-023-00818-0 |
| work_keys_str_mv | AT saadshsammen rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT ozgurkisi rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT mohammadehteram rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT ahmedelshafie rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT nadhiralansari rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT mohammadalighorbani rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT shakeelahmadbhat rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT alinajahahmed rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms AT shamsuddinshahid rainfallmodelingusingtwodifferentneuralnetworksimprovedbymetaheuristicalgorithms |