| 摘要 |
The present invention relates to a method for representing a T-box, in which a class that is an object or a an objective, properties, and properties between the class and a class are defined, with bit cubes, and an area of an A-box defined by an instance based on a concept defined in the T-box, with a bit matrix and more particularly, to an inference search method using a hyper cube comprising an improved method for building a hyper cube index. The initial method for building a hyper cube index comprises the steps of: i) changing rows and columns of index data; ii) binding the length N of the row into a number L in the index data of an N * N matrix (array) having only bit values of 0 and 1 and representing a binary number using a decimal number to compress into index data of an N * (N / L) matrix (array); iii) processing N iteration and data compression operations in parallel according to an N triple number when generating index data; and iv) dividing the index data of the N * (N / L) matrix (array) into K rows to store index data of a K * (N / L) matrix (array). The method for building a hyper cube index according to the present invention can generate a hyper cube index within a reasonable time even in a large ontology, and can process interference queries within a practical time. |
