摘要 |
<p>A system compares a target case against a database of records, responsive to proximity between that target case and the database's records, returning records with highest proximity. Proximity is responsive to how many attributes match and how many attributes differ. Each attribute has an attribute value representing an index into that attribute's information. The searchable records are ordered by number of attributes: Records with few attributes precede those with many attributes. They are further ordered lexicographically by attribute value: Records with smaller attributes precede those having only larger attributes, as described herein. This allows rapid elimination of records whose proximity cannot match the least-best records thus found. Records are examined for attributes which most contribute to increased proximity, allowing search specificity improvement by updating those attributes.</p> |