MACHINE LEARNING IN THE PREDICTION OF PARAMETERS OF THE HEAVY RAINS EQUATION. CASE STUDY: NORTHWEST FLUMINENSE

Abstract

The heavy rain forecasts are important in disaster prevention as well as agricultural planning. In this paper parameters of the heavy rainfall equation were estimated using a machine learning model applying the random forests technique and with a case study in the region of Noroeste Fluminense, Brazil. The same parameters were also estimated, using the Levenberg-Marquardt optimization method, using data from satellite images. Both the Machine Learning parameters predicted by the Levenberg-Marquard model are compared with the parameters found in the literature.