| 摘要 |
<p>A method of data search in a system for the automatic processing of information, using the vocal recognition of alphabetic type requests, provides for the preliminary identification of a set of candidate records which are ordered for each required field, according to the application of an algorithm for the management of the uncertainty in the search, which is of a probabilistic type. A certain number of record candidates is therefore chosen, with the best probabilities, for a cross search on the same database. A device is described, which carries out the method. <IMAGE></p> |
