Stream-Masters of Science in GIS & Remote Sensing Prediction of Flood by Using Artificial Intelligence in the Case of Dire Dawa Watershed, Awash Basin, Ethiopia
| dc.contributor.advisor | Binyam Tesfaw | |
| dc.contributor.author | Habte Kinfe | |
| dc.date.accessioned | 2026-07-28T11:17:11Z | |
| dc.date.available | 2026-07-28T11:17:11Z | |
| dc.date.issued | 2022-12 | |
| dc.description.abstract | In the present century, floods are one of the foremost destroying and costly risks around the world. Its contribution can be ascribed to variables like climate change impacts on the hydrologic cycle, land-use changes and increased density of residence activities in flood-prone areas. However, Flood is a complex event to model and predict for the future. Recently Machine Learning/ Artificial Neural Network (ML/ANN) enhanced the accuracy of weather related event prediction. In Dire Dawa, Ethiopia, flood is forecasted by conventional methods and this method is less accurate and cannot predict the flood hazard effectively. Therefore, this study aims to improve flood prediction accuracy in the area by using Artificial Intelligence(AI) such as; SVM (Support Vector Machines), MLP (Multi-layer perceptron),and KNN (KNeighborsClassifier). In this study, data were downloaded from Copernicus European environmental Reanalysis data Agency (ERA5), preprocessed and trained by using python language. Hourly precipitation and runoff data for 31 years were used for the training of algorithms. Out of these data, 75% were used for training and 25% for testing and validation while doing the machine learning process. The result shows that SVM was found to be the robust performing algorithm with Mean Squared Error of 0.0005046, and R2 of 99.949% compared to MLP (99.947%) and KNN (99.939 %). These methods are useful for the prediction and recommended for the policy makers to consider the AI method for the flood prone areas of the country. | |
| dc.identifier.uri | https://etd.aau.edu.et/handle/123456789/8709 | |
| dc.language.iso | en_US | |
| dc.publisher | Addis Ababa University | |
| dc.subject | Flood | |
| dc.subject | Flood Prediction | |
| dc.subject | Remote Sensing | |
| dc.subject | Dire Dawa | |
| dc.subject | Artificial Intelligence | |
| dc.title | Stream-Masters of Science in GIS & Remote Sensing Prediction of Flood by Using Artificial Intelligence in the Case of Dire Dawa Watershed, Awash Basin, Ethiopia | |
| dc.type | Thesis |