Matching engine with a signature generation and relevance

摘要

A kind of system and a kind of method generate at least one signature associated with document. In one embodiment, the document being made of text and parsing are received, with token group. The token set includes multiple tokens. Each token corresponds to the text separated in document by predefined character feature. Score calculates in document of each token token set based on frequency and distribution text. Each token is then according to score calculated. The sequence dry food selection signature of subset generates token selected by each appearance. Selected book single catalog simultaneously exports.