Hashing of Databases with the Use of Metric Properties of the Hamming Space

摘要: Hashing of databases is considered from the point view information and coding theory. The records a database are represented as binary vectors same length stored in external memory computer. task formulated follows: given pattern fixed size working memory, form set addresses that can disagree with numberof positions smaller than threshold value. We use metric properties Hamming space show computational efforts needed to search for be essentially decreased by using triangle inequality distances between vectors. Furthermore, an introduction Lee distance containing leads new where effectively used.