Association rules for solar flare attributes through hierarchical clustering

Abstract

This project aims to explore a set of association rules through hierarchical clustering as a way to preprocess the rules. The goal is to have a limited set of association rules based on the creation of groups which would indicate promising rules. The data used in the clustering and the construction of the rules are parameters of solar flares. For the construction of a proper data set of solar flares, the following actions were done i) choose a set of data; ii) preprocess the data in order to transform categorical data into binary data; iii) create groups by hierarchical clustering; iv) select groups to create association rules; v) apply the rules using an association algorithm; vi) validate the association rules.