作者: Michael D. Gordon
DOI: 10.1016/0306-4573(88)90100-8
关键词:
摘要: Abstract A document retrieval system may be described by three formal characteristics: the syntax employed to describe documents (keywords or vectors of weights, for instance), form machine-processable queries it accepts as valid (unordered sets keywords, keywords with Boolean connectives weighted vectors, example), and rules used rank retrieve documents. This article argues that interdependence among descriptions, queries, requires adaptation perform effectively when one its components changes. Recently, suggestions have been made modify traditional systems allow more flexible ranked output. However, these new forms likely require differently than they are in existing, commercial systems. “genetic algorithm” is discussed a means redescribing probabilistic algorithm uses feedback along alternative descriptions single takes account dependency structure subject terms.