FUZZY CLASSIFICATION OF SKEWED DISTRIBUTIONS: DETECTING SEVERE UPDRAFTS

V LAKSHMANAN1

ARTHUR WITT 2

National Severe Storms Laboratory
Norman, OK 73069, USA

Abstract:

When candidate signatures are disproportionately distributed, using the standard fuzzy operators [5] for aggregation could lead to many false alarms. We introduce a symmetric fuzzy aggregation operator that depends on the ordering of its arguments as well as on their values. We show that, using this aggregation operator, we can classify candidate updraft signatures with greater skill than we could by using standard fuzzy operators.