The weighted median is introduced as a fusion operation which can be u
sed in situations in which, while having numeric values for the weight
s associated with the objects to be fused, the actual objects being fu
sed only satisfy an ordering property. After introducing the concept o
f weighted median we compare it with the weighted average and show tha
t they have many properties in common. We then provide an algorithm fo
r learning the weights associated with a median aggregation. We then s
how how we can use this technique to extend the applicability of the O
rdered Weighted Averaging (OWA) operator to situations in which the ar
guments are nonnumeric. Finally we show how we can use the weighted me
dian as an alternative to the expected value in the evaluation of prob
abilistic lotteries. (C) 1998 Elsevier Science Inc.