Modelo predictivo de ocurrencia de eventos extremos de la variabilidad climática en las costas ecuatorianas basado en la detección de patrones de distribución de organismos zooplanctónicos utilizando minería de datos.

Abstract

The present work "Predictive model of occurrence of extreme events of climatic variability in Ecuadorian coasts based on the detection of distribution patterns of zooplanktonic organisms using data mining" aims to obtain a model capable of predicting the occurrence of extreme climatic events such as those of the ENSO (El Niño Southern Oscillation) cycle based on the behavior of zooplanktonic organisms, through the use of data mining techniques, the large volume of data used belongs to the permanent program of Climate Variability of the National Fisheries Institute, who has a historical database on the subject. The data mining techniques were executed through the use of the KDD (Knowledge Discovery in Database) methodology, using the KMINE tool, which is a technological platform specialized in the execution of data mining algorithms , such as: J48 and PRISM, which are algorithms that belong to data mining of descriptive type, for which the type of algorithms used are those of classification, such as those mentioned above. The ENSO cycle is a topic of great interest, because its occurrence brings with it great consequences such as: floods, scarcity, abundant rainfall, caused by the phenomena that belong to this cycle. A predictive model is a tool for decision making and for the elaboration of plans that maintain the integrity of the entire population. The proposed predictive model allows to determine the occurrence of ENSO Cycle events such as: El Niño, La Niña or Normal Conditions, which through relevant factors taken from the database after the mining process and analysis, is generated through rules obtained through the algorithms used.