Construction and mining of Object Cubes

In an object-oriented databases,classes of objects rather than individual objects are generalized and multidimensional analysis are applied on them.Generalization of large set of objects becomes difficult because these objects share attributes and methods of other classes.

In order to perform a class-based generalization on a large set of objects,the extended version of the attribute -oriented induction method is used.

In case of large data sets,the generalization can be seen as an application of class-based data mining process performed on different attributes in sequence.This generalization can be obtained efficiently by analyzing each attribute,generalizing them into a simple-valued data and then constructing a multidimensional data cube called an object cube.Data Mining and multidimensional analysis can be performed on these object cubes in a similar way as that of relational data cubes.However,it is very difficult to generalize the data set into a single-valued data.For example,consider an attribute keyboard which consists of a set of keywords.These keywords are used to define the features of the book.Because of which,it is not sensible to generate these set of key words into single-valued keyword.

0 comments:

Post a Comment