Floods are one of the worst natural disasters in the world. Colombia is a country that has been greatly affected by this disaster. For example, in the years 2010 and 2011 there was a heavy rainy season, which caused floods that affected at least two million people and there were economic losses of 6.5 million dollars, which is equivalent to 5.7% of the country's Gross Domestic Product (GDP) at that time. The Magdalena River is the most important since 128 municipalities and 43 cities with a population of 6.3 million people, which is 13% of the total population of the country, are located in its basins. For this reason, the objective of the research is to design and implement a model that helps predict flooding over the Magdalena River by examining three techniques of artificial intelligence (Artificial Neuronal Networks, Adaptive Neuro Fuzzy Inference System, Support Vector Machine), and thus determining which of these techniques are the most effective according to the case study. The research was limited only to these three types, due to limitations of time, data, human and financial resources, and technological infrastructure. In the end, it is concluded that the Artificial Neural Networks technique is a suitable option to implement the predictive system as long as it is not very complex and does not require high processing machine. However, to establish a model based on rules to achieve a better interpretability of the floods, the ANFIS model can be used.
Moreno Jenny Marcela, Sánchez Juan Manuel, Espitia Helbert Eduardo
Artificial Neural Networks, Artificial intelligence, Climate variability, Climatology, Computer engineering, Control systems, Earth sciences, Environmental economics, Environmental science, Flood, Magdalena River, Neuro Fuzzy Systems, Support Vector Machine, Systems engineering