Retrieval language is an artificial language created according to the needs of literature retrieval, also known as retrieval logo. From the point of view of reflecting the characteristics of literature, the retrieval marks that represent the external characteristics of literature, such as author's name, title, report number, standard number, patent number, file number, classification number, description, title and keywords, are all retrieval languages. However, from the perspective of standardization of retrieval marks, retrieval languages can be divided into natural language retrieval marks and standardized language retrieval marks. When compiling retrieval tools, indexers should analyze all kinds of documents, analyze all the content points contained in them, make them form some concepts that can represent the content of documents, and mark these concepts with standardized languages such as descriptors, topic nouns or classification numbers, and incorporate them into the retrieval system. When searching, users should analyze the topic of the question, form concepts that can represent information needs, and then convert these concepts into languages acceptable to the system, and then they can get documents indexed in these standardized languages from the system. Therefore, transforming the natural language of information demanders into a systematic and standardized retrieval language has a great relationship with the success of retrieval.

At present, information retrieval languages are mainly divided into two categories: 1, systematic classification and classified retrieval languages; 2. Topic method and topic retrieval language.

For more complex retrieval, it is best to use several retrieval languages to retrieve from different ways, and each retrieval language has its own advantages and disadvantages. System classification language has one-dimensional characteristics, which is suitable for family retrieval by subject system, but not for multi-dimensional feature retrieval by subject concept. Subject language, whether descriptive or headline, has the advantages of directness, specificity and flexibility, which overcomes the shortcoming that systematic classification can only retrieve documents from one concept, and the lack of national retrieval ability becomes its shortcoming. Although the thesaurus uses inverted titles or a large number of references to concentrate the search marks with internal relations, it still cannot overcome the contradiction of scattered similar documents and affect the recall rate. In addition, the pre-group nature of the title also determines its lack of ability to describe complex concepts. Thesaurus language is developed on the basis of absorbing the advantages of many languages, and its advantages of flexible combination are mainly reflected in computer retrieval. Manual retrieval is rarely used, and it is not as good as classified language system. In a word, if we fully understand the advantages and limitations of the above retrieval languages, we can foster strengths and avoid weaknesses, which is of great benefit to improving the precision and recall.

withdraw