Database system in which logical principles for a data retrieval process can evolve based upon learning

Abstract

The present invention provides a database system capable of dealing with dynamic changes in an event that vary continuously with time. A database system outputs a retrieval output from data in response to a user's retrieval request on the basis of a retrieval logic. The retrieval logic is allowed to evolve on the basis of a logical model. This logical model establishes a general relationship which leads results from conditions relating to a group of events that are dealt by a data group. Moreover, a database system optimizes a data structure on the basis of the logical model. The logical model can be transferred in a separated manner from the specific data contents of the data group. Therefore, the logical model can be exported to another database system of the same kind, and also imported from another data base system of the same kind.

1 . A database system comprising: 
 a data store for storing a data group;    a logical model holding unit for holding a logical model which represents logical principles in a data retrieval process in an independent manner from respective data contents of said data group;    a retrieval logic applying unit which creates a retrieval logic corresponding to a retrieval request on the basis of said logical model, and obtains a retrieval output corresponding to said retrieval request by retrieving said data group in accordance with said retrieval logic;    a logic analyzing unit which analyzes a logical path connecting conditions to corresponding results in a group of events relating to said data group; and    a logical model updating unit which updates said logical model so as to be evolved, each time a new combination between new conditions and corresponding new results is given, on the basis of an analysis of a logical path of the new combination carried out by said logical analyzing unit.    
     
     
         2 - 18 . (canceled)