Affiliation:
1. University of Petroleum and Energy Studies, India
2. Nanyang Technological University, Singapore
3. Kyungpook National University, South Korea
Abstract
The government has focused to maintain the needs of the populace's health and hygienic standards; numerous initiatives are involved, such as flood forecasting, water management, and sewage management. To prevent damage throughout the city, flood prediction must be done early on. “Smart” refers to artificial intelligence or machine learning methods, either directly or indirectly. To comprehend the general pattern and depth of the rainfall and to forecast the occurrence of floods, artificial intelligence techniques like deep learning are applied. To extract key properties for forecasting heavy rains and floods, many deep learning approaches, including CNN and deep belief networks, are applied. As a result, there is less harm done to both city infrastructure and human life. The study done on flood forecasting utilizing AI, ML, and deep learning techniques will be covered in this chapter. This review research will provide a thorough analysis based on the many types of deep learning models, the input datatypes for forecasting, the model effectiveness, real-time application, etc.
Reference41 articles.
1. Bankruptcy forecasting: An empirical comparison of AdaBoost and neural networks
2. An “Internet of Things” vision of the flood monitoring problem.;M.Ancona;The Fifth International Conference on Ambient Computing, Applications, Services and Technologies,2015
3. Boosting Regression Estimators
4. Smart flood disaster prediction system using IoT & neural networks
5. Brochero, D., Anctil, F., & Gagné, C. (2012, April). Forward Greedy ANN input selection in a stacked framework with Adaboost. RT-A streamflow forecasting case study exploiting radar rainfall estimates. In EGU General Assembly Conference Abstracts (p. 6683). ADS.