Boosting of Association Rules for Robust Emergency Detection

摘要

The use of association rules extracted from daily geophysical measures allows for the detection of previously unknown connections between events, including emergency conditions. While these rules imply that the presence of a given symbol occurs while a second one is present, their classification performance may vary with respect to test data. We propose to build strong classifiers out of simpler association rules: their use shows promising results with respect to their accuracy.