Flood Forecasting Using Weather Parameters

Many of us have experienced flooding in Malaysia. It has several negative effects on our way of life, including shortage of food, mortality, and property destruction. To reduce the impact of flood disasters on our daily lives, flood forecasting is important. The weather warning is vital to safeguarding people and property. Despite the ability to analyze climate change due to the rapid development of geographic information systems and the availability of data from several sources, data mining applications are scattered and systematic efforts on climate data mining are limited. Thus, our proposed system aims to provide insight into weather changes. The system will be able to predict floods by using weather parameters such as rainfall and temperature. Recurrent Neural Networks (RNN) can be applied to these parameters in order to predict future weather. RNN is an efficient method because it can handle sequential data and accept both the input data being used now and inputs from the past as the internal memory allows them to remember prior inputs. This system can have easier access to the rainfall conditions ahead. Besides, Predictions of extreme weather changes or natural disasters might be used to implement preventive measures.