JPH01112431A - Method for constituting hierarchical knowledge system - Google Patents

Abstract

PURPOSE:To access an efficient knowledge by providing functions to search a network composed of a node and an arc and to execute the pattern matching. CONSTITUTION:In a node and arc network 1, a node 6 desires to access the information given from a node 15 by the succession. In this case, in order to obtain the information, it is necessary for a searching and pattern matching function 2 of a system to arrive at a node 15 through an arc. Once the searching is executed from the node 6 up to 15, as the result, the information that nodes 9, 10, 13 and 15 are the master nodes of the node 6 is accumulated as the mapping information. Thus, in the second access of knowledge, even when the searching and pattern matching operation to require a complicated operation and a time is not executed, the knowledge can be accessed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for configuring a hierarchical knowledge system, which is a fundamental component of knowledge information processing.

[Conventional technology]

Conventional technology in this field is, for example, published patent publication No. 61.
-502707.

[Problem that the invention seeks to solve]

Such knowledge systems are common and important. However, these hierarchical knowledge systems require a huge amount of pattern matching processing and search processing to access knowledge. Therefore, the efficiency of the system will be greatly influenced by how the system is actually configured. The present invention provides an efficient method for configuring this system.

[Means for solving problems]

In the configuration of a hierarchical knowledge system that stores knowledge as a network consisting of node names, node attribute names, node attribute values, and arcs connecting nodes, the present invention provides at least a) a network consisting of nodes and arcs; (b) includes in its configuration a mechanism for detecting changes in the network as a result of knowledge access in the network; and C) orders nodes into their multiple parent nodes. This is a method for configuring a hierarchical knowledge system, characterized in that the configuration includes a 1:n mapping function for mapping, and d) a function for erasing and modifying mapping information of the node.

〔Example〕

FIG. 1 shows a conceptual diagram of an embodiment of the present invention. The number l in the figure indicates a network consisting of nodes and arcs, 2 indicates a search and pattern matching function for this network, 3 indicates a function for accessing this network and detecting network changes due to it, and 4 indicates a node. 5 shows a 1:n mapping function for mapping the data to its plurality of parent nodes, and 5 shows a function for erasing and modifying this mapping information.

The method of accessing knowledge in this configuration method will be explained using FIG. 2 as an example.

In the node/arc network shown in FIG. 2, it is assumed that node 6 wants to access information given from node 15 through inheritance. In this case, the search and pattern matching functions of the system need to reach the node 15 through the arc to obtain that information. -Finally, this operation requires a considerable amount of time.

In the configuration of this system, if the -degree nodes 6 to 15 are searched, the result will be nodes 9 and 15.
．． Information that 13.15 is the parent node of node 6 is stored as mapping information.

Therefore, when accessing such knowledge for the second time, it becomes possible to access the knowledge without performing search or pattern matching operations that require complicated operations and time.

Furthermore, since knowledge access through such mapping depends on the structure of the node and arc networks, when the network changes as a result of knowledge access, this mapping information also changes. Therefore, for example, node 1 in Figure 2
If 0 is deleted, the mapping information of node 6 to the parent node will be deleted and modified to only node 9.

〔Effect of the invention〕

As is clear from this embodiment, the method for configuring a hierarchical knowledge system according to the present invention allows knowledge to be accessed much more efficiently than in the conventional case.

[Brief explanation of the drawing]

FIG. 1 is a conceptual diagram of an embodiment of the present invention. The number l in the figure indicates a network consisting of nodes and arcs, 2 indicates the exploration and pattern matching function of this network, 3 indicates the function of accessing this network and detecting network changes due to it, and 4 indicates the node. A 1-to-n mapping function for mapping to a plurality of parent nodes is shown, and 5 shows a function for erasing and modifying this mapping information. FIG. 2 is a conceptual illustration of the operation of the hierarchical knowledge system according to the present invention. Number 6.7.8.9.1O111S12.13.
14 and 15 each indicate a node. Applicant: Seiko Epson Corporation Figure 2

Claims (1)

[Scope of Claims] A method for configuring a hierarchical knowledge system that stores knowledge as a network consisting of node names, node attribute names, node attribute values, and arcs connecting nodes, comprising: at least a) nodes and arcs; b) includes in its configuration a mechanism for detecting changes in the network as a result of knowledge access in the network, and c) connects the node to its multiple parent nodes. 1. A method for configuring a hierarchical knowledge system, characterized in that the configuration includes a 1:n mapping function for sequentially mapping to the node, and d) a function for erasing and correcting mapping information of the node.