Abstract
An approach to objects or events similarity is based on the similarity of the data values of the specific attributes. Similarity is refined by considering importance weights for attributes and also the issues of unusual attribute values where the concept of importance amplification is used to provide soft matching of objects or events We then introduce extensions to hypermatching where certain combinations of attributes are relevant. This is approached by modeling how to represent commonly occurring attribute data values whose co-occurrence is uncommon. Certainly not all attribute combinations are typically of the same interest. What can be expected is that for a particular context or application, some subset of the attributes is being focused upon. As an application, we illustrate the importance of considering combinations of attribute values in assessing evidence in geospatial profiling.