Entity resolution system identifying non-distinct names in a set of names

摘要

A method for identifying non-distinct names in a set of names (e.g. people, buildings, places, organizations, documents, cars, tings, objects etc.) comprises obtaining the set of names for a first entity and in response to comparing a first name and a second name in the set of names, it is determined that the first name is similar to the second name. Initials in the first name and the second name are searched for. In response to the search indicating that there is at least one initial in at least one of the first name and the second name, it is determined that the at least one initial matches a corresponding initial in another one of the first name and the second name and one of the first name and the second name are marked as a non-distinct name. The remaining names are considered to be distinct names to be used for scoring against another set of names for a second entity using a cross-entity scoring technique. Thus, the entity resolution system finds the most distinct name(s) among a group of related names for a single entity before attempting to resolve possibly related entities.