Knowledge Processing Technology: Interconnection of Heterogeneous Databases by Correlation Measurement based on Semantic Space Model

Abstract

In this paper, we will introduce the new technology of correlation search a retrieval technique based on measurement space model. There are said to be two types of human’s knowledge. One is the kind we obtain from reading literature. This may be called absolute knowledge. With ICT technology, it is knowledge that can be described by the ontology used in a conceptual dictionary and Semantic Web. Another kind may be called relative knowledge. This is paradigmatically, relational knowledge we gain when we compare knowledge A with knowledge B. Such knowledge may be connected in thought without being displayed (i.e, as a document on the Internet or being inscribed on a piece of paper). Humans often compare heterogeneous things. Moreover, humans are thought to unconsciously create the criteria for linking heterogeneous things. In this model, the criteria are represented by a set of axes which form space and by a measurement method which mathematically measures norms, distance or inner product. It is important to search not only for the correlation between heterogeneous knowledge, but also for factors (axis, index) that contribute to the correlation. This model will show correlations between heterogeneous knowledge and their factors, by deriving a selected knowledge cluster that has high correlation magnitude and a set of axes used to calculate the magnitude.