Structured Prior Knowledge and Granular Structures - Brain Informatics - page 124

Table 1. Training dataset

U Education Vocation Income(unit:yuan)

1 Doctoral student Private enterprise High

2 Postgraduate student State-owned enterprise High

3 Others Education Low

4 Undergraduate Private enterprise Low

5 Undergraduate State-owned enterprise Low

6 Postgraduate student State-owned enterprise Low

7 Undergraduate State-owned enterprise High

8 Undergraduate Civil servant Low

9 Doctoral student Education Low

10 Others State-owned enterprise Low

Fig. 6. Multidimensional data model
of table 1

Fig. 5. Granular structure of table 1

4 Relation between Structured Prior Knowledge and

Granular Structure

Every one may be possess much knowledge, which usually be organized as a
complicated network architecture in human brain. But prior knowledge is always
related to a problem. When we solve a problem, we only need to consider prior
knowledge relevant to it. Prior knowledge relevant to the solved problem will
be extracted from the complicated network structure, and be reorganized as a
nested and hierarchical structure, we call it structured prior knowledge, which
will helpful to human problem solving.
Granular structure is a central notion of granular computing, it can repre-
sents a problem from multilevel and multiview. Granular structure can make im-
plicit knowledge explicit, make invisible knowledge visible, make domain-specific
knowledge domain-independent and make subconscious effects conscious [7]. In
essence, granular structure is a structure hid in the solved problem.
As stated above, we can see that prior knowledge and granular structure are
related together tightly by the solved problem. There have many points of sim-
ilarity between prior knowledge and granular structure. For instance, they have