Thes-Maker: A Tool for Legal
The main purpose of an information retrieval system is to supply, on request, all the pertinent information stored in the system and only that information. There are, therefore, two requirements: to retrieve all the relevant information (recall capacity) and to exclude all irrelevant information (precision capacity).
Information retrieval is based on a technique that compares two sets of linguistic "signs" used as languages: the indexing language, on the one hand, which expresses the information, and the query language on the other hand, which expresses the request for information.
The difficulty in retrieval lies in making the search or query language correspond to the language of the data bank (or indexing language), and where these two do not correspond the performance of the system declines, giving rise to silence and/or noise.
The fact that almost all information retrieval systems use natural language (pure or corrected) at the indexing level creates enormous problems in retrieval and increases the difficulty relating to correspondence mentioned earlier. Among the traditionally more efficacious tools in improving quality in the selection of information, at both the level of recall and precision, thesauri play a prominent role.
A thesaurus can be defined as a structured lexicon. It consists of a set of words called the vocabulary and a set of relations between the individual words or groups of words making up the vocabulary.