Machine learning for river flow modeling and flood prediction

Abstract

The object of the research are predictive mathematical models or prediction algorithms applied to data provided by sensor technology, radio frequency identification (RFID) and internet of things (IOT). These algorithms will be used in the monitoring of environmental variables with wireless sensor networks connected to the Internet of Things and prediction of hazards and natural disasters based on advanced tools and methods of Data Science including, but not limited to artificial neural network models and machine learning algorithms integrated with Big Data Analytics platforms.For intelligent cities, the prediction paradigm is of paramount importance because it is pointed out as the last step in its modernization whose goals, among other advances, are the control and prevention of natural disasters.Therefore, the central objective of this research is to identify the mathematical models with the best assertiveness rates and that result in early flood diagnosis.In addition, we will analyze the ways in which the data of the sensors are collected and how these data are related in order to decode the behavior of these natural phenomena as well as to identify the reasons that make a given model more assertive in relation to the too much. And, finally, to evaluate the impacts of this process of diagnostics in the so-called smart cities. (AU)