|
Title:
|
ANALYSIS OF EXTREME EVENTS USING A DATA MINING APPROACH |
|
Author(s):
|
Heloisa M. Ruivo, Haroldo F. de Campos Velho, Fernando M. Ramos, Saulo R. Freitas |
|
ISBN:
|
978-989-8533-45-6 |
|
Editors:
|
Hans Weghorn |
|
Year:
|
2015 |
|
Edition:
|
Single |
|
Keywords:
|
Data mining, statistical analysis, t-test, p-value, decision tree, Shannon entropy. |
|
Type:
|
Full Paper |
|
First Page:
|
113 |
|
Last Page:
|
120 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The increasing volume of data in the environment sciences is a challenge for analysis and interpretation. Among the difficulties generated by this data deluge is the development of efficient knowledge discovery strategies. Here, we apply methods from statistics and computational intelligence to analyze large data sets of climate science. These techniques are simple and robust, and generate a mapping becoming easier the interpretation. Our approach comprises two steps for knowledge extraction. The first step applies a statistical method for class comparison. The second step consists of a Decision Tree (DT) classifier, based on learning algorithm. The DT is used as a predictive model to identify the precipitation intensity. The methodology is used to identify and for understanding extreme rainfall events. The method is employed to identify the more significant meteorological variables associated to the event. The technique is applied to two extreme precipitation events occurred in Brazil: Santa Catarina state (2008), and another one in Rio de Janeiro state (2010). |
|
|
|
|
|
|
